**** BEGIN LOGGING AT Sun Dec 19 02:59:57 2010 Dec 19 03:25:28 build #47 of ramips is complete: Failure [failed shell_13] Build details are at http://tksite.gotdns.org:8010/builders/ramips/builds/47 Dec 19 03:56:23 jow * r24701 /packages/Xorg/lib/cairomm/patches/ (. 001-missing-macros.patch): [packages] cairomm: supply missing macros Dec 19 04:20:14 swalker * r24702 /packages/utils/nano/Makefile: [packages] nano: update to 2.2.6 Dec 19 04:26:03 swalker * r24703 /packages/lang/lsqlite3/Makefile: [packages] lsqlite3: update to 7 Dec 19 04:39:21 jow * r24704 /packages/libs/ (libiconv/ libiconv-full/ libiconv-full/Makefile): [packages] move libiconv to libiconv-full Dec 19 04:39:58 jow * r24705 /packages/libs/libiconv/ (8 files in 4 dirs): [packages] add stub libiconv Dec 19 04:40:37 jow * r24706 /packages/libs/gettext/src/m4/iconv.m4: [packages] gettext: remove iconv.m4, its now provided by libiconv (or full gettext) Dec 19 04:41:49 jow * r24707 /packages/libs/glib2/Makefile: [packages] glib2: let configure autodetect the iconv implementation type Dec 19 04:42:49 jow * r24708 /trunk/include/autotools.mk: [include] autotools.mk: like libintl, make all libtool fixup packages depend on the libiconv stub Dec 19 04:46:02 ... now I only need to figure out a way to switch between stubs and full implementation depending on the language support flag in menuconfig Dec 19 04:53:08 jow * r24709 /packages/libs/elfutils/Makefile: [packages] elfutils: remove libintl hacks from makefile Dec 19 04:54:09 jow * r24710 /packages/net/sshfs/Makefile: [packages] sshfs: remove libintl hacks from makefile Dec 19 05:11:05 jow * r24711 /packages/libs/libiconv/src/include/iconv.h: [packages] libiconv: add stddef.h include to stub header Dec 19 05:18:09 jow * r24712 /packages/libs/gettext/src/include/libintl.h: [packages] gettext: some programs expect libintl.h to include locale.h, so do it. Spotted in minidlna Dec 19 05:19:17 jow * r24713 /packages/multimedia/minidlna/patches/002-makefile-tweaks.patch: [packages] minidlna: remove explicit -lintl from patch and let configure figure it out Dec 19 05:21:18 jow * r24714 /branches/backfire/include/autotools.mk: [backfire] let the libtool fixup introduce libltdl and libiconv build depends, as in trunk Dec 19 05:22:08 build #42 of ixp4xx is complete: Failure [failed compile_7] Build details are at http://tksite.gotdns.org:8010/builders/ixp4xx/builds/42 Dec 19 05:32:06 nunojpg * r24715 /packages/utils/wifitoggle/ (4 files in 3 dirs): [packages] wifitoggle: small makefile update Dec 19 05:46:08 nunojpg * r24716 /packages/net/sshtunnel/files/initd_sshtunnel: [packages] sshtunnel: fix typo Dec 19 06:23:50 nunojpg * r24717 /packages/utils/wifitoggle/files/ (hotplug-wifitoggle hotplug_wifitoggle): [packages] wifitoggle: filename change Dec 19 07:31:09 xMff: iptables is still no closer to being fixed Dec 19 09:06:11 build #41 of au1000 is complete: Exception [exception failed slave lost shell_13 compile_12] Build details are at http://tksite.gotdns.org:8010/builders/au1000/builds/41 Dec 19 09:20:12 xMff, acoul, whoever else: r21795 seems to cause #8211; what to do? Dec 19 09:40:59 xMff: btw, a rebuild after a distclean fixed my earlier problem libXau for me Dec 19 09:42:37 also, btw, /me enjoying a few sips of izarra, anybody from france? Dec 19 10:07:27 juhosg * r24718 /trunk/target/linux/ar71xx/ (3 files in 3 dirs): ar71xx: fix micrel phy patch (Patch from #8041.) Dec 19 10:46:06 sn9: so netfilter.mk is out of sync after all Dec 19 10:46:29 oh geez, acoul sure knows how to REALLY break things (r24345) Dec 19 10:46:46 #8366 Dec 19 10:46:58 breaking things++ Dec 19 10:47:18 xMff: i fixed netfilter.mk, but #8366 persists Dec 19 10:59:35 the layer7 patches places the files in iptables.../ not in iptables.../extensions/ Dec 19 10:59:38 this looks odd Dec 19 10:59:58 it completely wrong Dec 19 11:01:44 http://paste.openwrt.org/d36a617b9 Dec 19 11:03:02 did it even compile a libxt_layer7.so ? Dec 19 11:12:16 apart from that I do not even have the required symbols in my kernel config Dec 19 11:12:18 no idea why Dec 19 11:18:18 jow * r24719 /trunk/package/iptables/patches/002-layer7_2.22.patch: [package] iptables: fix layer7 patch to place the source in extensions/, I doubt it ever worked Dec 19 11:45:18 hauke * r24720 /trunk/tools/e2fsprogs/patches/ (2 files): Dec 19 11:45:18 e2fsprogs: add two compile fixes. Dec 19 11:45:18 Thank you Joe Roback for reporting and fixing compiling on Darwin. Dec 19 11:50:03 nbd * r24721 /trunk/package/hostapd/files/hostapd.sh: (log message trimmed) Dec 19 11:50:04 hostapd: Allow rsn_preauth configuration Dec 19 11:50:04 Hostapd can allow preauthentication for WPA2-EAP networks when Dec 19 11:50:04 the interfaces through which preauthentication is allowed are Dec 19 11:50:04 configured. Dec 19 11:50:04 Add a new param "rsn_preauth=0/1" to the configuration that enables Dec 19 11:50:05 or disables preauthentication on the according bridge interface. Dec 19 11:50:11 nbd * r24722 /trunk/package/hostapd/files/hostapd-full.config: Dec 19 11:50:11 hostapd: Add CONFIG_FULL_DYNAMIC_VLAN=y to hostapd-full.config Dec 19 11:50:11 This allows hostapd to make use of the config option Dec 19 11:50:11 "vlan_tagged_interface" to automatically create bridge and Dec 19 11:50:11 VLAN interfaces in a fully dynamic way. Dec 19 11:50:12 Signed-off-by: Helmut Schaa Dec 19 11:58:14 nbd * r24723 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-ubnt.c: Dec 19 11:58:14 ar71xx: enable ar7240 switch support on the UBNT NanoStation M Dec 19 11:58:14 Fixes broadcast frames on the second ethernet MAC Dec 19 11:58:14 Signed-off-by: Alexander Couzens Dec 19 11:58:22 nbd * r24724 /trunk/package/hostapd/files/hostapd.sh: Dec 19 11:58:22 hostapd: fix appending the iapp_interface variable to the config Dec 19 11:58:22 Signed-off-by: Alexander Couzens Dec 19 11:58:28 nbd * r24725 /trunk/package/mac80211/files/lib/wifi/mac80211.sh: Dec 19 11:58:28 hostapd: add support for specifying the maximum number of clients Dec 19 11:58:28 Signed-off-by: Alexander Couzens Dec 19 12:09:36 nbd: the tl-wr[89]41nd also still use the dsa switch stuff according to grep Dec 19 12:10:52 yeah, but iirc that one doesn't use the ar724x dsa stuff Dec 19 12:11:10 at least one of those is marvell Dec 19 12:13:46 ah, ok Dec 19 12:28:16 xMff or anyone: this fixes #8211 and #8366: http://www.gimpelevich.san-francisco.ca.us/danielg4/r24725.patch Dec 19 12:30:33 also, the p2pblock luci pkgs are missing a dependency on the ipp2p iptables module, and likewise, the ipp2p kernel module is missing a dependency on compat_xtables Dec 19 12:31:51 nbd: so now, who's actually maintaining brcm-wl? Dec 19 12:32:31 well, nobody's actively maintaining it Dec 19 12:33:50 what was the point of negotiating it to be gpl'ed if there is still no source? Dec 19 12:34:11 what do you mean by 'negotiating it to be gpl'ed'? Dec 19 12:34:41 broadcom did release a GPL'd mac80211 driver Dec 19 12:34:55 oh Dec 19 12:35:00 with ap? Dec 19 12:35:01 but it only supports a limited range of hardware Dec 19 12:35:04 no AP yet Dec 19 12:36:23 well, on mimo/11n hardware, brcm-wl appears not to be functional. on 2.4, i could at least have unencrypted ap, but no more Dec 19 12:37:38 my guess is that the problem is somewhere in the open source parts or in the hackish ssb integration Dec 19 12:37:54 since the binary part of the driver was built from pretty much the same source code that dd-wrt uses Dec 19 12:38:02 and dd-wrt supposedly works on such hardware Dec 19 12:38:31 is openwrt's wl that different from dd-wrt's wl, which has fully supported 11n for years? afaik, nothing like that ever worked on openwrt Dec 19 12:38:51 the open parts are different Dec 19 12:38:57 the closed parts not so much Dec 19 12:39:34 the main changes to the open parts were related to hooking into ssb Dec 19 12:39:45 and fixing bugs with newer kernel releases Dec 19 12:39:51 dd-wrt is still on 2.4, and the wl that openwrt had on 2.4 didn't use the open ssb Dec 19 12:39:57 no, dd-wrt is on 2.6 as well Dec 19 12:40:07 but not using ssb Dec 19 12:40:23 dd-wrt is using 2.6.23 Dec 19 12:40:42 latest service pack from dd-wrt is still 2.4; i checked their svn' Dec 19 12:40:50 they have both 2.4 and 2.6 Dec 19 12:40:59 but 2.6 is the default Dec 19 12:41:21 the beta of the next service pack has 2.6 Dec 19 12:41:47 there is no clear versioning in dd-wrt Dec 19 12:41:51 everybody just uses the beta Dec 19 12:41:52 ;) Dec 19 12:42:42 but if the closed part is the same between the 2.4 dd-wrt and the 2.4 openwrt, why did one fully support 11n so long and the other never? Dec 19 12:43:21 or was the difference purely in userspace? Dec 19 12:43:25 because nobody in openwrt cares enough about broadcom to do proper testing and fix the remaining bugs Dec 19 12:43:49 the 2.6 driver in openwrt is pretty much the same as the 2.6 driver in dd-wrt Dec 19 12:43:55 i don't know much about the similarity on the 2.4 side Dec 19 12:43:59 nor do i care about it Dec 19 12:44:12 the 2.6 driver should be fixed, but i don't have time to do that Dec 19 12:44:23 i opened up as much of the code as i could without violating the licence Dec 19 12:44:41 in hope that somebody else would do the rest of the work, as i have too much work to do on ath9k and other things Dec 19 12:45:15 i have no experience with dd-wrt on 2.6, but i know what dd-wrt did on 2.4 and what openwrt did on 2.4 and does on 2.6 Dec 19 12:46:01 i know that all of brainslayer's broadcom builds are 2.6 (the 2.4 builds are done by another guy) Dec 19 12:46:08 and he tests primarily on 11n hardware Dec 19 12:46:14 so i'm pretty sure it works to some degree there Dec 19 12:47:01 dd-wrt on 2.4 fully supported all the 11n features, openwrt on 2.4 worked only at 11g speeds and only unencrypted, and openwrt on 2.6 just plain doesn't work Dec 19 12:47:21 yes, you've mentioned that already Dec 19 12:47:52 i am wondering how this came about Dec 19 12:48:10 probably the user space side on 2.4 wasn't finished either Dec 19 12:48:21 i never tested with anything bug legacy hw Dec 19 12:48:28 s/bug/but/ Dec 19 12:48:47 if the driver is essentially the same, does this mean that all it needs is userspace fixing? Dec 19 12:49:01 i think you're getting things mixed up again Dec 19 12:49:06 the 2.4 driver is quite different from the 2.6 driver Dec 19 12:49:21 and the 2.4 driver in openwrt is older than the 2.4 driver in dd-wrt Dec 19 12:49:35 likely to have more bugs Dec 19 12:50:07 the 2.6 driver probably has a chance of working if somebody fixes the open source parts Dec 19 12:50:19 i don't think the bugs in the 2.6 driver are primarily user space issues Dec 19 12:50:27 i would assume the 2.6 version of dd-wrt supports everything the 2.4 one did, no? Dec 19 12:50:44 if it was missing user space stuff, the driver should have worked in legacy Dec 19 12:50:56 well, probably not everything, there may be some older hardware that doesn't work in 2.6 anymore Dec 19 12:51:14 but the 2.6 driver also supports some newer hardware that didn't work with 2.4 Dec 19 12:51:19 i meant as far as 11n hardware goes Dec 19 12:51:55 maybe Dec 19 12:52:05 but i don't care about the 2.4 driver at all Dec 19 12:52:11 looking into it will not help fix the 2.6 driver Dec 19 12:52:59 so, if the closed parts of the 2.6 driver in owrt and dd-wrt are the same version, there is a common reference point Dec 19 12:53:08 everything necessary for fixing the 2.6 driver should be in the open parts of the trunk brcm-wl package and whatever you can salvage from the dd-wrt repository Dec 19 12:53:36 obviously i can't rule out bugs in the closed parts, but i think bugs in the open parts are far more likely at this point Dec 19 12:54:23 well, if the ssb stuff is to blame, wouldn't one at least need to know what the closed parts expect of the open ssb code? Dec 19 12:54:54 the API and the broadcom sb implementation are open Dec 19 12:55:24 in fact ssb only does the basic init, then from the driver's perspective the broadcom ssb support code takes over Dec 19 12:55:29 and that part is still open Dec 19 12:55:44 it's compiled along with the brcm-wl package and included as source code Dec 19 12:56:10 and isn't that part essentially the same in dd-wrt now? Dec 19 12:56:20 no Dec 19 12:56:24 as i said, dd-wrt does not use ssb Dec 19 12:56:43 so it doesn't have the tweaks that i did to make the broadcom sb stuff work together with ssb Dec 19 12:56:57 most of the low level code is the same, yes Dec 19 12:57:09 but in dd-wrt the broadcom sb code is used from the arch code as well Dec 19 12:57:40 so in openwrt there might be some conflict between the upstream ssb code and the broadcom sb (siutils) code Dec 19 12:58:32 but isn't that so low-level that a problem there would cause a more basic failure? Dec 19 12:59:11 well, not necessarily Dec 19 12:59:19 it may be that ssb simply does one simple aspect differently Dec 19 12:59:21 as it stands, it creates the ap, but nothing can associate, triggering immediate timeouts Dec 19 12:59:25 e.g. a different clock setting Dec 19 12:59:32 or some other setting that only affects the 802.11 core Dec 19 12:59:35 hmm Dec 19 12:59:51 the 11n router that you're testing with - does it have the wireless part behind pci or as a wisoc? Dec 19 13:00:27 ssb: PCIcore in host mode found Dec 19 13:00:27 ssb: CardBus slot detected Dec 19 13:00:51 that doesn't answer my question Dec 19 13:01:20 the aforementioned cardbus holds the wifi Dec 19 13:01:36 wifi behind cardbus? that's odd Dec 19 13:02:18 the card doesn't even fit the cardbus formfactor; only the slot does Dec 19 13:03:03 well, it kinda sorta fits... Dec 19 13:04:07 so maybe ssb cardbus slot handling is broken Dec 19 13:04:21 or some other thing related to that Dec 19 13:05:05 occasionally it sees the cardbus card as having the pci id 0000:4000 and a power cycle fixes it Dec 19 13:07:32 but it handles it well enough to create the ap otherwise Dec 19 13:08:14 yeah, but maybe it has dma issues Dec 19 13:08:33 sounds plausible Dec 19 13:11:11 have there been any reports of wl working on ANY 11n hardware on 2.6 owrt? Dec 19 13:11:32 i don't know, i haven't really followed it Dec 19 13:12:24 but it was verified on 11g, right? Dec 19 13:12:48 yes, at least on some hardware Dec 19 13:12:54 there are some hardware revisions that do not work Dec 19 13:13:13 specifically lp-phy Dec 19 13:13:25 *headdeask* Dec 19 13:13:31 -a Dec 19 13:13:47 this 11n has a lp-phy Dec 19 13:14:01 lp-phy is not 11n lp-phy Dec 19 13:14:08 oh Dec 19 13:14:56 what's known about the issue with 11g lp-phy? Dec 19 13:15:19 not much Dec 19 13:17:07 does it manifest the same way? Dec 19 13:17:36 seems like Dec 19 13:18:19 probably the same issue Dec 19 13:20:56 have you seen the lp-phy issue yourself? Dec 19 13:21:21 didn't have time to look into it properly Dec 19 13:21:37 no hardware? Dec 19 13:22:50 right, at least not the same that this problem was reproduced on Dec 19 13:23:46 any data you'd like to collect from my hardware? Dec 19 13:25:51 no, i don't know what to look for Dec 19 13:26:03 brb Dec 19 13:26:30 anything you'd like to try? Dec 19 13:27:33 btw, someone really should commit that patch i pasted above; broken iptables on a router is kinda silly Dec 19 13:35:28 jow * r24726 /trunk/package/mac80211/files/lib/wifi/mac80211.sh: [package] mac80211: rename max_clients option to maxassoc, this brings it in line with broadcom Dec 19 13:36:15 xMff: hi Dec 19 13:38:12 hi Dec 19 13:38:56 jow_laptop: did you see the patch? Dec 19 13:39:14 yes, but it needs to be cleaned up first Dec 19 13:40:05 you can quilt refresh that part Dec 19 13:43:05 whitespace looks messed up too Dec 19 13:43:33 anyway, I'm online with gprs right now Dec 19 13:43:38 it will take a while Dec 19 13:44:04 * sn9 is online with hspa Dec 19 13:46:02 there are also the dependency issues i mentioned Dec 19 13:51:39 well Dec 19 13:51:46 not a luci issue Dec 19 14:29:29 dingo * r24727 /packages/lang/ruby/Makefile: [patch-team] - Upgrade ruby to 1.9.2-p0 - corrects file clashes between ruby-readline and ruby-irb - corrects file clashes between ruby-webrick and ruby-openssl - Signed-off-by: Vincent Neyen Dec 19 15:38:42 Does anyone know the link to this IRC log that one can browse through? I left this IRC around 10:15PM last night and would like to view what had been on the IRC since then. Dec 19 15:44:59 build #47 of rb532 is complete: Failure [failed shell_9] Build details are at http://tksite.gotdns.org:8010/builders/rb532/builds/47 Dec 19 16:20:34 why does the wiki main page have a chinese image builder explanation since almost a year? Dec 19 16:20:43 mazilo: http://logs.nslu2-linux.org/livelogs/openwrt-devel.txt Dec 19 16:21:39 KanjiMonster: link? Dec 19 16:22:22 [Sun 2010-12-19 04:28:55 AM PST] also, the p2pblock luci pkgs are missing a dependency on the ipp2p iptables module, and likewise, the ipp2p kernel module is missing a dependency on compat_xtables Dec 19 16:22:25 xMff: go to the main page and scroll down Dec 19 16:22:32 also visible here: http://wiki.openwrt.org/doc/howto/imagebuilder Dec 19 16:22:55 that is certainly not since a year Dec 19 16:22:57 has a chinese header first, then goes into the english stuff.. has a link to the chinese and english versions in the top right Dec 19 16:23:03 more likely < 2 weeks Dec 19 16:23:47 fixed Dec 19 16:23:55 been since here: http://wiki.openwrt.org/doc/howto/imagebuilder?rev=1292641371 Dec 19 16:24:44 it was there for a day Dec 19 16:25:36 xMff: hrm, okay, then is "old revision" somewhat broken - I used it to look at old revisions, there it was in most of them Dec 19 16:25:55 yiu probably used the diff icon Dec 19 16:26:02 it is always based against most recent Dec 19 16:26:09 so not really useful Dec 19 16:26:46 I clicked on "old revision" an then clicked myself through the oder ones till I found one where it wasn't Dec 19 16:26:56 KanjiMonster: well that's what i just did and found the one above Dec 19 16:26:58 worked fine... Dec 19 16:27:13 anyway, enough chatter about a 15 sec wiki revert Dec 19 16:27:32 yeah Dec 19 16:27:58 we could need more wiki mods Dec 19 16:28:21 this does not imply writing stuff, but the permission to move, rename and delete pages Dec 19 16:28:51 actually we have a few but often ppl. are mia after a few days Dec 19 16:29:05 is there a mechanism that alerts you to changes and such? Dec 19 16:29:23 there's an rss feed Dec 19 16:31:43 http://wiki.openwrt.org/feed.php Dec 19 16:31:59 well, there's a bit too much going on for me right now, but after christmas i can probably help.. Dec 19 16:31:59 xMff: ah, I only looked at the reivisions of /start; didn't notice its an included page - no wonder I didn't see it in the revision log Dec 19 16:32:36 ahh :) Dec 19 16:32:48 [Sun 2010-12-19 08:25:41 AM PST] anyway, enough chatter about a 15 sec wiki revert Dec 19 16:35:06 xMff: still cleaning up my patch? Dec 19 16:35:10 sn9: Thanks for the link, but I am looking for the IRC log of yesterday to see if there was any response from xMff w.r.t. PHP5 fix. Dec 19 16:40:03 xMff: Ah I see you are now online. Sorry that I left earlier last night and won't know if you had the patches on PHP5 for me to test. Dec 19 16:40:35 mazilo: I committed everything and now wait for the buildbot results Dec 19 16:40:50 xMff: OK. Dec 19 16:41:47 everything? there is still my patch for #8211 and #8366 Dec 19 16:48:21 jow * r24728 /trunk/target/linux/generic/ (3 files in 3 dirs): [generic] fix layer7 patches for Kernels >= 2.6.35, thanks Daniel Gimpelevich Dec 19 16:49:03 jow * r24729 /trunk/include/netfilter.mk: [include] netfilter.mk: fix connmark packaging for Kernels >= 2.6.35, thanks Daniel Gimpelevich Dec 19 16:53:17 Kaloz: ping Dec 19 16:58:55 xMff: i discovered two missing deps too, one in luci and one in xtables-addons Dec 19 17:03:37 build #52 of ppc40x is complete: Failure [failed shell_9] Build details are at http://tksite.gotdns.org:8010/builders/ppc40x/builds/52 Dec 19 17:05:12 freifunk-p2pblock should depend on iptables-mod-ipp2p, and kmod-ipt-ipp2p should depend on kmod-ipt-compat-xtables Dec 19 17:08:02 nbd: is there anything you'd like to try with my 11n hardware to shed light on possible lp-phy issues Dec 19 17:08:06 ? Dec 19 17:10:51 sn9: u on a rage ?! Dec 19 17:11:55 blogic: if that means "in a hurry to leave" then yes Dec 19 17:23:11 sn9: as i said, i don't really have time to look into this Dec 19 17:23:34 i'll take that as a "no" Dec 19 17:23:35 i'd have to make myself more familiar with the hardware to come up with some useful tests Dec 19 18:08:39 swalker * r24730 /packages/admin/htop/Makefile: [packages] htop: fixup the autotools mess for backfire (#8486), thanks jow Dec 19 18:13:23 build #42 of kirkwood is complete: Failure [failed shell_9] Build details are at http://tksite.gotdns.org:8010/builders/kirkwood/builds/42 Dec 19 18:18:53 updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/uscan.shtml Dec 19 18:29:40 xMff: Is result from built #42 you have been waiting for? Dec 19 19:15:13 build #51 of ubicom32 is complete: Failure [failed compile_4] Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/51 Dec 19 20:23:19 WARNING: can't parse line: USB_IP_COMMON=m Dec 19 20:23:20 WARNING: can't parse line: USB_IP_HOST=m Dec 19 20:23:20 WARNING: can't parse line: USB_IP_VHCI_HCD=m Dec 19 20:24:23 just noticed that fly by in a brcm47xx compile Dec 19 20:28:36 jow * r24731 /packages/net/usbip/Makefile: [packages] usbip: fix bad kconfig symbols, thanks russell Dec 19 21:58:27 loswillios: ping Dec 19 22:05:49 mirko * r24732 /packages/sound/pulseaudio/Makefile: [packages/pusleaudio] fix (host)dependencies (intltool) Dec 20 01:15:24 jow * r24733 /packages/libs/gettext/ (Makefile src/include/libintl.h): [packages] gettext: remove duplicate defines, thanks hlu **** ENDING LOGGING AT Mon Dec 20 02:59:58 2010