**** BEGIN LOGGING AT Mon Feb 02 02:59:57 2009 Feb 02 03:43:24 lol Feb 02 04:13:59 svn: OPTIONS of 'https://svn.ntop.org/svn/ntop/trunk/n2n': Server certificate verification failed: issuer is not trusted (https://svn.ntop.org) Feb 02 06:53:14 chrispitzer: if you have two wireless interfaces, yes Feb 02 06:53:26 TWO PHYSICAL INTERFACES Feb 02 07:19:22 russell_: tarball uploaded, that should be resolved. Feb 02 07:19:48 kicking off a run to see if we are missing any other source tarballs. Feb 02 07:20:09 doing it manually once seemed to fix it too. Feb 02 07:20:38 seemed to be a 'accept this self-signed cert' thing Feb 02 07:21:11 yeah, we run svn with the --no-interactive flag, so untrusted self-signed certs cause failure. Feb 02 07:41:06 nico * r14351 /trunk/include/quilt.mk: prevent quilt from loading any configuration file (closes: #4520) Feb 02 08:09:10 hmm. build failure for ar71xx: Feb 02 08:09:14 cp: cannot stat `/aux/src/openwrt/build_dir/linux-ar71xx/linux-2.6.27.13/drivers/net/phy/libphy.ko': No such file or directory Feb 02 08:09:14 make[3]: *** [/aux/src/openwrt/bin/packages/target-mips_uClibc-0.9.29/kmod-tg3_2.6.27.13-ar71xx-1_mips.ipk] Error 1 Feb 02 08:09:14 make[3]: Leaving directory `/aux/src/openwrt/package/kernel' Feb 02 08:40:03 <[florian]> russell_: which driver are you building ? Feb 02 08:40:10 <[florian]> russell_: did you select tg3 ? Feb 02 13:39:35 nico * r14352 /packages/libs/zaptel-1.4.x/Makefile: [packages] zaptel-1.4.x: split builds of kernel & userland stuff, use $(CP) instead of $(INSTALL_DATA) for shared library, bump release number Feb 02 14:24:09 mirko * r14353 /feeds/phone/paroli/ (5 files in 3 dirs): uprade paroli the next working version, removing everything but the dialer, modify naming and desktop-files Feb 02 14:37:10 florian * r14354 /trunk/package/kernel/modules/netdevices.mk: [package] package libphy separately and make tg3 depend on it for kernels 2.6.27 and superior Feb 02 15:36:00 mirko * r14355 /feeds/phone/paroli/files/paroli.cfg: adding missing config file Feb 02 15:49:20 nico * r14356 /trunk/package/kernel/modules/ (block.mk netfilter.mk): [package] kernel: mark modules that won't build on 2.6.28 as broken Feb 02 16:12:23 <{Nico}> TheCompWiz: he is not here right now, but he might show up later today Feb 02 16:12:34 k... what's his nick? Feb 02 16:13:26 <{Nico}> mb, mb_ or mb__ Feb 02 16:13:35 hm... shoulda guessed /slapforehead Feb 02 16:14:49 {Nico}: he has many face muhahaha Feb 02 16:18:34 does anyone in here use the broadcom hardware? (bcm5354-ish?) Feb 02 16:19:52 guess that's a no. Feb 02 16:39:33 nbd * r14357 /trunk/include/kernel-defaults.mk: strip the .note.gnu.build-id section from kernel images, which on some platforms can lead to unusable 3.1G kernel image files Feb 02 16:43:25 nbd * r14358 /trunk/toolchain/binutils/Config.in: make binutils 2.18 default for ppc44x and ppc40x Feb 02 16:45:18 nbd * r14359 /branches/8.09/ (include/kernel-defaults.mk toolchain/binutils/Config.in): merge r14357,r14358 to 8.09 Feb 02 16:48:24 jow * r14360 /branches/packages_8.09/ipv6/radvd/ (Makefile patches/003-linux_old_compat.patch): [packages_8.09] merge radvd fix from r14349 Feb 02 16:57:51 nbd * r14361 /packages/net/mini_snmpd/patches/102-compile_fix.patch: fix mini_snmpd compile on avr32 Feb 02 16:59:32 nbd * r14362 /branches/packages_8.09/net/mini_snmpd/patches/102-compile_fix.patch: merge r14361 to 8.09 Feb 02 17:24:33 juhosg * r14363 /trunk/target/linux/adm5120/files/arch/mips/adm5120/cellvision/nfs-101.c: [adm5120] nfs-101u: add PCI IRQ map for the NEC USB controller (thanks to Daniel Schmitt for testing) Feb 02 17:25:46 nbd * r14364 /packages/net/stun/ (Makefile patches/002-linking_order.patch): fix stun compile on avr32, brcm-2.4 Feb 02 17:26:41 nbd * r14365 /branches/packages_8.09/ (32 files in 16 dirs): sync some packages with trunk Feb 02 17:51:54 juhosg * r14366 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/mach-mzk-w04nu.c: [ar71xx] mzk-w04nu: add GPIO LEDs and buttons Feb 02 17:59:07 [florian]: it was a 'select all packages' build (or was originally ... in this case i was make oldconfig'ing an earlier select-all, and saying 'm' to everything i was asked) Feb 02 18:01:16 hmm. at r14366, i see this: Feb 02 18:01:26 mips-openwrt-linux-uclibc-gcc -c -Os -pipe -mips32r2 -mtune=mips32r2 -funit-at-a-time -fhonour-copts -I. -g -fPIC -Wall -DBUILDING_TONEZONE -m64 -DSTANDALONE_ZAPATA -DZAPTEL_CONFIG=\"/etc/zaptel.conf\" -o zonedata.lo zonedata.c Feb 02 18:01:27 cc1: error: unrecognized command line option "-m64" Feb 02 18:01:27 make[4]: *** [zonedata.lo] Error 1 Feb 02 18:18:51 nbd * r14367 /trunk/toolchain/binutils/Config.in: default to binutils 2.18 on ppc* for manual selection as well Feb 02 18:21:30 nbd * r14368 /branches/8.09/toolchain/binutils/Config.in: merge 14367 to 8.09 Feb 02 18:25:48 nbd * r14369 /branches/8.09/target/linux/ppc44x/Makefile: ppc44x should no longer be broken Feb 02 18:59:38 {Nico}: zaptel problem Feb 02 19:01:34 <{Nico}> russell_: what's wrong? Feb 02 19:01:38 <{Nico}> pastebin? Feb 02 19:01:51 {Nico}: when built on x64 a -m64 is leaking in Feb 02 19:01:59 x86_64 that is Feb 02 19:02:56 see build_dir/linux-ar71xx/zaptel-1.4.9.2/Makefile, look for uname -m Feb 02 19:03:41 <{Nico}> russell_: ah ok Feb 02 19:03:53 * {Nico} needs an x86_64 build machine too Feb 02 19:06:24 if i manually hack that part out of the makefile, a la: Feb 02 19:06:28 #ifneq (,$(findstring x86_64,$(UNAME_M))) Feb 02 19:06:28 #CFLAGS_x86_64:=-m64 Feb 02 19:06:28 #endif Feb 02 19:06:33 it builds okay Feb 02 19:06:56 it's easier than that Feb 02 19:07:08 that's because i don't know what i'm doing Feb 02 19:07:14 add UNAME_M=$(ARCH) to the make command Feb 02 19:07:19 in the build system makefile Feb 02 19:07:45 it overrides the makefile's own variable assignment Feb 02 19:07:52 thus doesn't need patching Feb 02 19:07:59 parfait Feb 02 19:10:24 oops. maybe not. Feb 02 19:11:53 ? Feb 02 19:12:15 hold on, testing Feb 02 19:12:31 yeah, first try i was being dumb Feb 02 19:12:36 second try better Feb 02 19:12:41 ok Feb 02 19:14:48 define Build/Compile/user Feb 02 19:14:48 $(MAKE) -C $(PKG_BUILD_DIR) \ Feb 02 19:14:48 UNAME_M=$(ARCH) \ Feb 02 19:14:56 that fixes the build problem Feb 02 19:15:13 + UNAME_M=$(ARCH) \ Feb 02 19:24:05 nico * r14370 /trunk/toolchain/gcc/patches/3.4.6/820-i386_inhibit_libc.patch: [toolchain] gcc: add a patch to allow 3.4.6 to build for x86 (closes: #4548) Feb 02 19:38:00 juhosg * r14371 /trunk/target/linux/ar71xx/files/arch/mips/ar71xx/platform.c: [ar71xx] parse dot-separated MAC address (patch by Jonas ) Feb 02 19:39:37 can someone tell me if there's an easy way to find out the arguments passed from cfe to the kernel? Feb 02 19:42:26 TheCompWiz: fw_argX Feb 02 19:42:56 I know that... but not sure how to actually get the values from a running device... Feb 02 19:43:04 TheCompWiz: cat /proc/cmdline Feb 02 19:43:15 actually mb__ you're the guy I've been looking for! Feb 02 19:43:18 LOL Feb 02 19:43:25 :) Feb 02 19:43:46 Well, from a running system you could probably load a module that prints the args Feb 02 19:43:57 I was just about to try & track down what's up with the lzma-loader & bcm 5354 CPU... and everyone says you're the man to talk to on that front. Feb 02 19:44:37 I'm currently using a custom CFE and the standard openwrt image. Feb 02 19:44:47 No lzma stuff Feb 02 19:45:08 ugg... ok. I don't suppose you know whats going on then... Feb 02 19:45:13 * TheCompWiz sighs. Feb 02 19:45:28 I never got openwrt booting with the original CFE, btw Feb 02 19:45:39 It's just really really really broken Feb 02 19:45:55 I have it working now... just as long as I comment out the init_cfe & console stuff... Feb 02 19:46:09 http://bu3sch.de/gitweb?p=cfe.git;a=summary I'm running this fork of CFE Feb 02 19:46:51 oh ok. Well. CFE changes the way it passes the args all the time. I think they just want to upset you. :) Feb 02 19:46:52 nico * r14372 /trunk/target/linux/generic-2.6/ (config-2.6.26 config-2.6.27): [kernel] generic: sanitize config files (fix typos & whitespace, remove duplicates) Feb 02 19:47:34 TheCompWiz: There is no working universal way to detect the CFE args. Feb 02 19:47:45 They all break in one or the other way. Feb 02 19:47:55 I think it would be best to disable CFE console support by default. Feb 02 19:48:14 You hardly need it anyway. Feb 02 19:48:22 I kinda like having cfe console support... (no jtag here) Feb 02 19:48:26 ssb serial is initialized very early Feb 02 19:49:22 before ssb serial init there are only 4 or 5 lines of kernel log anyway Feb 02 19:49:26 is all that stuff *only* for initializing the serial console after cfe hands off to the kernel? Feb 02 19:49:57 CFE console support is _only_ for debugging the ssb serial stuff, IMO. It's no good for other stuff. Feb 02 19:50:12 After ssb serial was inited, it will take over the console. Feb 02 19:50:39 so, what the heck is with all the effort in setting up the early console? Feb 02 19:51:01 Well, dunno. I don't need it, because ssb serial simply works Feb 02 19:51:23 I guess it can be useful, if there goes something wrong in the very early MIPS arch code, before we init SSB. Feb 02 19:52:17 So that's why I think it's best to default it to OFF. Feb 02 19:52:46 It's only useful for hardcore early debugging if something goes wrong (which is not the case for 5354) Feb 02 19:53:01 I figured it wasn't booting at all since after it hit the "starting program at blah" console went dead... Feb 02 19:53:07 But it can possibly break otherwise working setups, because CFE just passes args in a braindamaged way. Feb 02 19:53:12 guess it was booting just fine... Feb 02 19:53:55 I never got below that line with the original CFE, except if I boot an ELF kernel only. Feb 02 19:54:25 well... commenting out the early_console stuff... it boots fine... Feb 02 19:54:39 ok nice. Maybe that was the case for me then, too. Feb 02 19:54:48 Most likely Feb 02 19:54:54 very odd still. Feb 02 19:54:56 I never debugged it Feb 02 19:55:08 anyhow... how goes work on the wifi driver? Feb 02 19:55:16 I'm working on it. Feb 02 19:55:27 making good progress I hope? Feb 02 19:55:31 yeah Feb 02 19:55:39 need someone to test something? Feb 02 19:56:32 No it's still very early stuff. Feb 02 19:56:38 oki doki Feb 02 19:56:50 I first need to get the bulk of stuff out before I run a test. Feb 02 19:57:13 It's probably about 2k-3k lines of code missing Feb 02 19:57:22 LOL ouch. Feb 02 19:57:37 I sure wish bcm would open-source their drivers completely. Feb 02 20:00:07 mb * r14373 /trunk/target/linux/brcm47xx/patches-2.6.28/810-ssb-add-pmu-support.patch: bcm47xx: Implement 4312 and part of 4325 PLL init. Feb 02 20:04:42 mb * r14374 /trunk/target/linux/brcm47xx/patches-2.6.28/810-ssb-add-pmu-support.patch: bcm47xx: Rename all SSB_PLLRES_ to SSB_PMURES_ Feb 02 20:07:52 TheCompWiz: Does boot_wait work for you, btw? Feb 02 20:08:01 It never did with the original CFE for me Feb 02 20:21:23 mb__: my cfe has a auto-recovery thingie... I don't use boot_wait Feb 02 20:25:59 oh. Ok. I never tried that. I guess I'm better off with my own stuff anyway :) Feb 02 20:26:05 LOL Feb 02 20:26:11 I'd really like to get uboot working on it :) Feb 02 20:26:47 well... I'm trying to keep as stock-as-possible... Feb 02 20:27:08 well... almost-stock Feb 02 20:27:20 that's boring :P Feb 02 20:27:22 I am not a fan of asus' default firmware... Feb 02 20:27:34 I never booted that Feb 02 20:28:00 cu later :) Feb 02 20:28:02 mb__: I agree it keeps things boring... but easy to trouble-shoot when one out of a batch screws up. Feb 02 20:28:06 doh. Feb 02 22:13:26 nico * r14375 /trunk/package/openssl/patches/180-fix_link_segfault.patch: [package] openssl: add a patch to fix link segfault when using a glibc/eglibc based toolchain (closes: #4517), thanks to Luigi Mantellini Feb 02 23:45:35 florian * r14376 /trunk/toolchain/binutils/Makefile: [toolchain] libssp should also be configured in binutils Feb 03 00:10:46 FYI - there is a new patch for the iw command to show the 802.11n rates of the newer drivers https://lists.ath9k.org/pipermail/ath9k-devel/2009-February/001056.html Feb 03 00:40:22 lars * r14377 /trunk/target/linux/generic-2.6/patches-2.6.28/960-fix-serial_h-includes.patch: Feb 03 00:40:22 serial.h exports struct serial_rs485 which contains __32 fields to userspace so it Feb 03 00:40:22 has to export linux/types.h to userspace aswell. Feb 03 00:40:22 Fixes #4445 Feb 03 02:46:33 nico * r14378 /packages/utils/lsof/Makefile: [packages] lsof: fix build on glibc/eglibc based toolchain, use build kernel version instead of host one, pass optflags without overriding cflags, bump release number Feb 03 02:50:12 nico * r14379 /trunk/target/linux/generic-2.6/config-2.6.25: [kernel] generic: add missing PHY related config entries Feb 03 02:55:46 Bartman007: Ever get a chance to take a look at my mailer package? It replaces the buggy ssmtp that is aged **** ENDING LOGGING AT Tue Feb 03 02:59:57 2009