**** BEGIN LOGGING AT Mon Nov 24 02:59:57 2008 Nov 24 05:11:46 x86 for soekris has some kernel option damage (in the form of kconfig questions) as of r13332 Nov 24 05:36:26 oh. heh. the kernel build fails also. Nov 24 05:37:18 LD [M] fs/xfs/xfs.o Nov 24 05:37:18 make[5]: Leaving directory `/aux2/src/openwrt/build_dir/linux-x86_generic/linux-2.6.27.6' Nov 24 05:37:18 make[4]: *** [/aux2/src/openwrt/build_dir/linux-x86_generic/linux-2.6.27.6/.modules] Error 2 Nov 24 05:42:11 oh. wait. that's misleading because i was doing a make -j4. looks like the real problem is here: Nov 24 05:42:27 CC [M] crypto/ocf/crypto.o Nov 24 05:42:27 crypto/ocf/crypto.c: In function 'crypto_exit': Nov 24 05:42:27 crypto/ocf/crypto.c:1686: error: implicit declaration of function 'kill_proc' Nov 24 09:42:11 florian * r13335 /packages/net/nodogsplash/ (Makefile files/ files/nodogsplash.init): Install missing configuration file, htdocs and forward port the original initscript to the kamikaze rc.d scripts layout, bum release number (#4232) Nov 24 10:06:01 florian * r13336 /packages/ipv6/radvd/Makefile: Bump radvd to 1.2, cleanup makefile (#4230) Nov 24 11:03:29 juhosg * r13337 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/platform.c: [ar71xx] fix gpio-leds, and gpio-buttons platform data initialization Nov 24 12:42:23 florian * r13338 /trunk/package/kernel/modules/crypto.mk: Hardware crypto modules require CONFIG_CRYPTO_HW=y to be set (#3883), thanks Norbert Nov 24 12:43:14 Has anyone experimented with uclibc nptl recently ? Nov 24 12:56:15 AR7Port Patch platform/110-flash.patch does not apply (enforce with -f) Nov 24 12:56:21 in last svn trunk Nov 24 14:47:45 ejka, matteo last trunk wont build (ar7): Patch platform/110-flash.patch does not apply Nov 24 15:01:45 xMff: ist "local salt" automatisch nil oder muesste ich "local salt = nil" schreiben? Nov 24 15:02:34 automatisch Nov 24 15:03:33 ok danke Nov 24 15:05:44 Welche URL wuerdest Du fuer das challageSalt.xml vorshlagen momentan hab ich http://localhost:8080/luci/sauth/? Nov 24 15:14:09 mir egal Nov 24 15:17:51 k Nov 24 15:32:56 florian * r13339 /packages/lang/php5/Makefile: Add php5-exif module (#4188), bump release number Nov 24 16:56:42 juhosg * r13340 /trunk/target/linux/ar71xx/files/ (3 files in 2 dirs): [ar71xx] ethernet driver preparation for gigabit support Nov 24 17:34:42 x86 build seems broken on "Select all packages by default". a simple checkout compiles, but with "Select all" it fails while compiling the kernel in that crypto/ocf/crypto.c place i mentioned earlier. Nov 24 18:12:52 <[florian]> russell_: ocf is broken for all architectures on 2.6.27, we are fixing it Nov 24 18:16:05 [florian], in ar7 110-flash.patch does not apply Nov 24 18:16:16 [florian], it is causing build error Nov 24 18:16:32 [florian], who is working on it? Nov 24 18:16:41 <[florian]> DagMoller: which svn revision, what's the exact problem ... ? Nov 24 18:16:51 [florian], lastest trunk Nov 24 18:17:36 [florian], http://openwrt.pastebin.com/m231b98eb Nov 24 18:17:51 <[florian]> ok thanks Nov 24 18:17:56 after that, no more patchs are applyed Nov 24 18:18:46 <[florian]> ok, I am building a trunk for ar7 Nov 24 18:19:15 [florian], i think this http://openwrt.pastebin.com/m3876392 solve the problem... just replace the files Nov 24 18:20:35 <[florian]> ok, will check that later Nov 24 18:20:37 <[florian]> thanks Nov 24 18:28:27 [florian]: is there something specific i can disable to let the rest build? Nov 24 18:31:33 russel_: have you tried to disable ocf in kernel_menuconfig ? Nov 24 18:31:47 yeah, trying that now Nov 24 18:33:03 i'm also getting 8 kernel config queries in a "select all" build Nov 24 18:36:37 net/netfilter/xt_layer7.c:444: error: 'struct nf_conn' has no member named 'counters' Nov 24 18:37:54 russel_: http://www.nabble.com/Unable-to-compile-using-kernel-2.6.20-td13119895.html - first response Nov 24 18:39:09 curious that's not fixed in 2.6.27 Nov 24 18:42:37 hmm. connection tracking flow accounting is selected, as shown by make kernel_menuconfig Nov 24 19:54:28 blogic: Umm... you owe me an email :) Nov 24 19:54:51 (and I really could use some testers for hotplug2, because the only thing pending for 1.0 is testing and deciding upon the UCI stuff) Nov 24 20:00:03 ok Nov 24 20:04:00 xMff: Sessions kommen jetzt nach [sessionpath .. "/session-" .. id] , okay? Ich mein, so koennte ich die Challanges einfach nach [sessionpath .. "/challenge-" .. username] packen...? Nov 24 20:04:33 "jetzt" ? Nov 24 20:07:18 xMff: ich mein mit den einbauen der MD5-Auth, oder haettest Du einwaende? Nov 24 20:10:25 mir egal solange hinterher nich alles kaputt ist Nov 24 20:11:12 ich frag mich nur warum du jetzt auch eigene Session-IDs hast Nov 24 20:11:30 machst du die ganze Zeit Basic-Auth? Nov 24 20:18:33 anyone builds openwrt on linux/amd64 saw the following libtool-related error ? Nov 24 20:19:27 http://pastebin.ca/1266070 Nov 24 20:19:33 this is in libopkg Nov 24 20:20:41 somebody else had the same problem a week ago or so. i think [florian] had a solution back then. Nov 24 20:20:44 juhosg * r13341 /trunk/target/linux/generic-2.6/patches-2.6.27/973-ocf_2.6.27_fix.patch: [kernel] add experimental fix for OCF on 2.6.27.y (closes #4184) Nov 24 20:21:48 larsc: did he post it to any list or a forum ? or do i wait ? Nov 24 20:22:19 dsr: ping him Nov 24 21:03:01 xMff: Ich mach keine eigenen Session-IDs ich will nur die Challages in das selbe Verzeichnis wie die schon vorhandenen Session-IDs packen, da so sich die Verwaltungs-Funktionen des Session-IDs auch um die Challages kuemmern wuerden. Und "session-" will ich vor die Session-Files packen, damit das system nicht hacken lasst, im den das session-cookie z.B. auf "challange-root" gesetzt wird. Nov 24 21:05:19 hi :) Nov 24 21:05:26 hi Nov 24 21:05:54 your attention is highly appreciated , wrt ;-) Nov 24 21:07:10 x-alina: äh ja, mach einfach Nov 24 21:08:37 hi Nov 24 21:08:44 hi Nov 24 21:08:48 lol Nov 24 21:08:52 doh Nov 24 21:08:56 shut up wrt Nov 24 21:09:00 :) Nov 24 21:09:01 Hello everybody... Nov 24 21:09:08 hello Nov 24 21:11:45 I have a problem with building openwrt on my computer (I'm trying to set up a series of VMs for meshnet testing)... every attempt to build openwrt (both trunk and rc) leads to an error when the image is created. Nov 24 21:12:16 which one? Nov 24 21:12:47 http://pastebin.com/d6f450163 Nov 24 21:13:12 I tried an x86 image and some other ones (one broadcom and one atheros)... same error Nov 24 21:13:54 for creating cross connected virtual machines to test olsrd Nov 24 21:14:42 here is a larger part of the error: http://pastebin.com/d138f2e87 Nov 24 21:20:56 guys, we would need to re-create (if possible) the setting for the olsr bug which was discovered in Berlin Nov 24 21:21:06 nbd: Hi... Aaron (Aroscha) told me you could help me with my openwrt problem ? Nov 24 21:21:15 and one attempt is to rebuild it via virtual machines Nov 24 21:21:31 like to reconstruct the settings of the bug Nov 24 21:21:37 but henning is stuck with the x86 bug Nov 24 21:21:42 of compiling openwrt Nov 24 21:21:43 Henning_Rogge: what problem? Nov 24 21:21:53 just a sec Nov 24 21:22:23 trying to compile a x86 version of openwrt to get a lightweight image for a virtual machine... always hit a bug during "make"... Nov 24 21:22:27 try disabling jffs2 in menuconfig Nov 24 21:23:41 okay... *testing* Nov 24 21:24:49 okay, it's working now... thanks a lot... :) Nov 24 21:25:02 :) Nov 24 21:25:03 thx Nov 24 21:26:42 cu Nov 24 21:26:47 cu Nov 24 21:32:57 got a build failure in coreutils when compiling r13338 for wgt634u Nov 24 21:34:35 x-alina: wie ist das Ganze eigtl. gegen Replay-Attacken gesichert? Nov 24 21:35:07 http://pastebin.ca/1266127 Nov 24 21:37:46 xMff: In dem die "challage-" .. username Datei mit der Zufallszahl nach dem erfolgreichen Login geloescht wird und ausserdem nach z.B. 10 Sekunden ungueltig ist. Nov 24 21:38:06 lars * r13342 /packages/Xorg/xserver/xorg-xserver/Makefile: xorg depends on libpciaccess. Nov 24 21:38:37 dann hast du aber das selbe Problem wie wir jetzt grade das man einfach später die Session-ID abgreifen muss Nov 24 21:39:01 Wie meinst Du das? Nov 24 21:39:41 so wie ich es sage, MITM beim Login hast du zwar jetzt erfolgreich verhindert aber danach muss man nur das Sysauth-Cookie rausfischen Nov 24 21:42:52 blogic: Do you know what magic is needed to boot a linux kernel with the zyxel bootloader crap? Nov 24 21:43:13 Ja, aber das ist nur eine bestimmte Zeit gueltig. Challange ist natuerlich nicht so sicher wie HTTPS, aber so laesst sich wenichstens das Password nicht mehr ohne weiteres mitsniffen. Das selbe Password haben ja auch eventuell andere Router die dann auch bedroht waeren. Ich hab es dafuer gedacht, falls Du mal vergisst https:// vor die URL zu schreiben usw. Nov 24 21:46:08 So dass Du Dich dann einfach nur ausloggen musst (Session-Datei wird dann doch zerstoert oder?) und duerfte sehr unwarscheilich sein, das in der kurzen Zeit der Router kompromitiert wurde. Nov 24 21:48:01 x-alina: kennst du grendl-scan.com ? Nov 24 21:48:20 hat einen fuzzer fuer web apps drinnen Nov 24 21:48:47 evtl. hilfts :) Nov 24 21:52:12 florian * r13343 /trunk/target/linux/ar7/patches-2.6.26/110-flash.patch: Fix this nasty patch which prevented others from applying Nov 24 21:54:58 aroscha: Gegen was testet er genau? Nov 24 21:56:08 x-alina: sql injections, generelles fuzzing, java URL in injections, etc etc etc etc... Nov 24 21:56:10 total viel Nov 24 21:56:15 mir isses nur gerade aufgefallen Nov 24 21:56:34 eher klappt der busybox-httpd zusammen ;) Nov 24 21:57:08 hehe, nun das waere auch ein guter test Nov 24 21:57:39 aroscha: could you please have a look at http://sourceforge.net/tracker/index.php?func=detail&aid=2119159&group_id=117612&atid=681702 would be nice if this could be fixed for 8.09 as its kind of essential for several wireless communities using olsrd Nov 24 21:59:17 xMff: Wieder ne dumme/google-faule Frage ;/ : Gibs in Lua auch sowas wie continue in for-Schleifen in anderen Sprachen? Nov 24 21:59:28 x-alina: nein Nov 24 22:04:58 cyrusFF: ok Nov 24 22:05:27 cyrusFF: ahm... actually I was just debugging dyn_gw before. c.f. the berlin list Nov 24 22:06:27 the to many pings I could not verify that Nov 24 22:06:38 aroscha: sorry, berlin-wireless is 90% flamewar 10% topic, maybe i've just missed the post Nov 24 22:07:00 cyrusFF: :))) understandable. Yes, I often don't read it myself. no problem. Nov 24 22:07:15 aroscha: need to test this soon again, there were some more issues, the patch referenced above fixes just a segfault on startup, but there are hna announcement related problems too Nov 24 22:07:26 aroscha: perhaps an endianess issue Nov 24 22:07:30 ok! Nov 24 22:07:34 this could be the case Nov 24 22:07:50 question: is it possible to make a openwrt mipsel for qemu? Nov 24 22:08:29 aroscha: and the dyngw plugin does strange sanity checks on hna cidrs too, disallows things like 192.168.0.0/16 which I find rather strange Nov 24 22:08:55 xMff: when you have it in the olsrd.conf file? Nov 24 22:09:08 like if you write 192.168.0.0/16 in the conf file? Nov 24 22:09:09 aroscha: qemu should be possible, I use vdi images for myself Nov 24 22:09:13 cool Nov 24 22:09:14 aroscha: yes Nov 24 22:09:22 this might be a way to test endian bugs Nov 24 22:09:24 well... Nov 24 22:09:28 let's put it like that Nov 24 22:09:37 the code goes thru /proc/net/route Nov 24 22:09:45 btw... any progress on ff-compat-issue? that's a blocker for us now Nov 24 22:09:47 and reads the hex representation of the routing table Nov 24 22:10:21 so there might be an endianness issue Nov 24 22:14:28 ok Nov 24 22:19:32 xMff: if it is an endianness issue: Nov 24 22:19:46 linksys is mipsel, right? Nov 24 22:19:56 yes Nov 24 22:20:18 arm is big endian? Nov 24 22:20:22 i mean the FON? Nov 24 22:20:28 yeah Nov 24 22:20:47 atheros soc Nov 24 22:21:14 that means if it is an endian issue it will a) work on neither or b) work on one of them (the linksys), since x86 is also little endian (and it works on x86 for me) Nov 24 22:21:41 if i don't mix up the little and bigs now again :))) Nov 24 22:22:08 breaking both in different ways would be the most fun Nov 24 22:22:13 hehe Nov 24 22:22:19 true Nov 24 22:31:43 xMff: Noch ne dumme Frage: return gibt automatisch nil zurueck oder? Nov 24 22:32:27 so in etwa Nov 24 22:33:12 wie genau? Nov 24 22:33:29 (wills sauber programmieren) Nov 24 22:36:23 return; ist halt void Nov 24 22:36:53 local foo = bar() ergibt dann nil in foo Nov 24 22:37:04 wenn bar() nur return; macht Nov 24 22:39:46 more x86 build damage: http://pastebin.ca/1266178 Nov 24 22:43:30 "return nil" und "return" ergiebt das selbe, so das ich auch einfach "return" schreiben kann?! Nov 24 22:43:39 ja Nov 24 22:45:10 THX :) Nov 24 22:47:06 ciao 2 everybody, gn8! Nov 24 22:47:22 gut nacht! :) Nov 24 22:53:49 target/linux/generic-2.6/patches-2.6.27/200-sched_esfq.patch Nov 24 23:08:28 russell_: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=5ce2d488fe039ddd86a638496cf704df86c74eeb Nov 24 23:09:24 yeah, it looks like that broke esfq Nov 24 23:10:11 * russell_ trying with that patch removed Nov 24 23:10:30 looks as if its easy to fix though Nov 24 23:11:29 oh, yeah. Nov 24 23:11:32 actually. Nov 24 23:14:34 s/sch->dev/qdisc_dev(sch)/ Nov 25 02:33:19 Is there some magic way to get Redboot to accept a dollar sign as part of a kernel parameter? Every time I try and pass memmap=4096$0x3fff000 it interprets the $ as a linefeed **** ENDING LOGGING AT Tue Nov 25 02:59:57 2008