**** BEGIN LOGGING AT Sun Apr 04 02:59:58 2010 Apr 04 03:43:15 nico * r20690 /trunk/include/netfilter.mk: [cosmectic] include/netfilter.mk: move ebtables definitions at the end Apr 04 06:15:01 build #42 of cobalt is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/cobalt/builds/42 Apr 04 06:41:03 abyrne * r20691 /packages/ipv6/wide-dhcpv6/ (Makefile files/dhcp6c.conf files/dhcp6c.init): [packages] wide-dhcpv6: Add debug option to dhcp6c Apr 04 08:53:15 gmorning Apr 04 09:16:17 identify joda_bot <$_o%37J Apr 04 09:16:28 awww Apr 04 09:17:02 hehe :D Apr 04 09:17:03 wops Apr 04 09:20:00 stintel: it's changed :D Apr 04 09:20:21 but thx for the highlight I might have missed my bubu Apr 04 09:21:01 didn't really hilight but I guess you don't like having irc windows with unread activity :P Apr 04 09:22:40 stintel: how did you figure out I'm on windows ... well done ... you're right Apr 04 10:58:14 xMff: are you there? Apr 04 11:00:25 I'm currently at the broken "network.alias" WRT54GS router Apr 04 11:01:01 perhaps someone can point me to debugging the network config/ finding the root of the problem Apr 04 11:01:11 I'm currently at the broken "network.alias" WRT54GS router Apr 04 11:07:31 bbl Apr 04 11:16:48 cyrus * r20692 /trunk/package/cyassl/Makefile: cyassl: shouldn't depend on zlib when compiled with --without-zlib Apr 04 11:53:30 build #56 of ppc44x is complete: Success [build successful] Build details are at http://tksite.gotdns.org:8010/builders/ppc44x/builds/56 Apr 04 12:02:16 build #36 of pxcab is complete: Failure [failed compile_3] Build details are at http://tksite.gotdns.org:8010/builders/pxcab/builds/36 Apr 04 12:06:52 [florian]: ping Apr 04 12:35:09 nico * r20693 /trunk/include/netfilter.mk: Apr 04 12:35:09 netfilter: extension fixes (partially closes: #7045) Apr 04 12:35:09 * add missing xt_owner (2.6) Apr 04 12:35:09 * enable ipt_quota (2.4), disabled in [8499] is building fine with recent iptables Apr 04 12:35:09 * add missing ipt_nat_tftp (2.4) Apr 04 12:35:10 * add missing nf_nat_amanda (2.6) Apr 04 12:40:07 nico * r20694 /trunk/package/kernel/modules/netfilter.mk: netfilter: fix extension list in package descriptions (closes: #7045) Apr 04 12:47:54 nico * r20695 /branches/backfire/ (include/netfilter.mk package/kernel/modules/netfilter.mk): [backfire] netfilter: backport r20690, r20693 & r20694 Apr 04 12:55:28 mirko * r20696 /packages/Xorg/lib/wnck/Makefile: libwnck depends on libX11 Apr 04 12:59:30 mirko * r20697 /packages/Xorg/lib/qt4/patches/200-depend-on-decoration.patch: check for any enabled decoration, not just for the default one Apr 04 13:01:52 mirko * r20698 /trunk/target/linux/xburst/config-2.6.32: do not disable netfilter support for xburst target (some packages require netfilter support) Apr 04 13:42:41 sigh cant scroll up to read hilight :p Apr 04 13:42:48 damn iphone terminal Apr 04 13:47:50 hy what is this file ? http://downloads.openwrt.org/snapshots/trunk/ar71xx/openwrt-ar71xx-tl-wr1043nd-v1-squashfs-factory.bin Apr 04 13:48:13 i dont no what is factory, and i dont see the universal Apr 04 14:00:02 anyone? Apr 04 14:01:16 speeedfire: well its most likely the default factory image as out of the box Apr 04 14:03:15 but wheil makeing any week new image? Apr 04 14:40:09 nico * r20699 /packages/libs/libmad/Makefile: [packages] libmad: fix build failure on targets using -O2 optimizations (partially closes: #6974) Apr 04 14:51:13 nico * r20700 /packages/sound/streamripper/Makefile: [packages] streamripper: explicitely disable use of internal libmad (closes: #6974), use PKG_INSTALL, bump release number Apr 04 15:39:18 lars * r20701 /trunk/target/linux/xburst/files-2.6.32/sound/soc/codecs/jzcodec.c: [xburst] jzcodec: Only support 16bit formats for now Apr 04 15:42:44 nico * r20702 /packages/net/asterisk-1.6.x/patches/210-ppc_fix.patch: [packages] asterisk-1.6.x: remove obsolete ppc fix (closes: #7069) Apr 04 15:47:28 nico * r20703 /trunk/target/linux/x86/base-files/lib/upgrade/platform.sh: sysupgrade: fix typo in platform_do_upgrade() on x86 (closes: #7068), thanks to acinonyx Apr 04 15:49:36 nico * r20704 /branches/backfire/package/cyassl/Makefile: [backfire] package/cyassl: backport r20692 Apr 04 15:51:27 nico * r20705 /branches/backfire/target/linux/xburst/config-2.6.32: [backfire] target/xburst: backport r20698 (closes: #7012) Apr 04 15:54:15 nico * r20706 /branches/backfire/target/linux/x86/base-files/lib/upgrade/platform.sh: [backfire] sysupgrade: backport r20703 Apr 04 15:55:05 nico * r20707 /branches/backfire/target/linux/xburst/files-2.6.32/sound/soc/codecs/jzcodec.c: [backfire] target/xburst: backport r20701 Apr 04 16:50:39 hi, who is maintaining the dev.openwrt.org trac site? Apr 04 16:53:00 I'd like to fix an annoying warning because I prefer fixing issues to ignoring them (in my browser). The logo on the page is afaik the only resource not served via https ... could someone fix the link to point to a site local https resource? Apr 04 16:53:19 Then my firefox won't complain about a mixed security site Apr 04 17:10:05 <_trine> I have just flashed my wrt160nl and it has ssh running before a passwd is set Apr 04 17:10:19 normal Apr 04 17:10:28 <_trine> made from trunk Apr 04 17:10:38 <_trine> is it Apr 04 17:11:13 <_trine> I wasnt aware that it should was Apr 04 17:11:20 <_trine> ok then Apr 04 17:12:13 <_trine> webif normally asks you to set a passwd but it doesn't in this version Apr 04 17:12:43 ssh is always running, it just does not let you login without a root pwd set Apr 04 17:13:25 <_trine> ok but webif if there is no passwd set normally asks you to set one but it isn't right now Apr 04 17:13:36 <_trine> if I do first boot it will Apr 04 17:14:16 <_trine> it's not a big problem but it's not correct Apr 04 17:15:54 <_trine> I thought that before you had set a passwd if you tried to connect via ssh it always said network refused previously Apr 04 17:16:10 <_trine> I must be mistaken Apr 04 17:16:25 maybe Apr 04 17:17:00 <_trine> but now I get the login prompt Apr 04 17:17:16 <_trine> oh well Apr 04 17:19:33 xMff: ping? Apr 04 17:21:31 philipp64: pong Apr 04 17:21:56 ok, so trying to take the net48xx profile, and use that as a starting point for the net5501. Apr 04 17:24:52 philipp64: what chip does the net5501 have for IDE? Apr 04 17:25:19 brb Apr 04 17:26:42 I think I need CONFIG_MFGPT_TIMER, CONFIG_CS5535, CONFIG_PC8736x_GPIO, CONFIG_NSC_GPIO, CONFIG_CS5535_GPIO, CONFIG_SCx200_ACB, CONFIG_AT24, CONFIG_SENSORS_PC87360 ... Apr 04 17:27:16 oh, and CONFIG_MGEODE_LX, of course... Apr 04 17:30:05 I think best is to make a subtarget Apr 04 17:31:05 isn't that what putting a description in target/linux/x86/generic/profiles does? Apr 04 17:31:33 hmmm... forgot CONFIG_GEODE_WDT ... Apr 04 17:31:33 philipp64: it's only for packages Apr 04 17:31:50 ok, guess I haven't figured out the difference. Apr 04 17:31:51 philipp64: and kernel modules Apr 04 17:32:04 philipp64: not core kernel config symbols Apr 04 17:32:31 philipp64: a subtarget allows you to override the generic kernel config Apr 04 17:32:34 right, but don't each of those kmod targets imply certain kernel config state? Apr 04 17:32:54 philipp64: it is a bit more than a profile since its kernel changes can make packages incompatible Apr 04 17:33:00 ok, so where do the subtargets live, and what can I use as a starting place? Apr 04 17:33:15 you can use the xen subtarget as a starting point Apr 04 17:34:31 target/linux/x86/xen_domu ? Apr 04 17:34:34 philipp64: you also have to modify the main platform Makefile to add the subtarget dir Apr 04 17:34:40 <{Nico}> philipp64: or the new ep80579, might be less intrusive Apr 04 17:37:21 the ep80579/config-default doesn't seem to contain much. Apr 04 17:37:40 it jsut contains the deltas Apr 04 17:37:49 or better the overrides Apr 04 17:38:19 <{Nico}> philipp64: right, it just override defaults from ../config-defaults Apr 04 17:38:19 ok, so what do I need to diff against? Apr 04 17:38:29 compared to target/linux/x86/config-default which itself is just overrides for target/linux/generic-2.6/config-default Apr 04 17:39:31 so basically add all symbols that are required for this board and short out those already in one of the generic configs Apr 04 17:41:03 yeah, unfortunately the ordering of symbols seems to change from one version of linux to another, so you can't just diff the files... as they are ordered differently. sigh. Apr 04 17:43:15 nbd * r20708 /trunk/target/linux/ixp4xx/patches-2.6.32/600-skb_avoid_dmabounce.patch: ixp4xx: add back the patch for avoiding dma bounce buffers for network buffers, which was lost on the kernel upgrade to 2.6.32. fixes ath9k and improves other drivers' performance on boards with >64M ram Apr 04 17:45:04 nbd * r20709 /branches/backfire/target/linux/ixp4xx/patches-2.6.32/600-skb_avoid_dmabounce.patch: [backfire] ixp4xx: backport r20708 Apr 04 17:47:07 xMff: does all the builtin hardware have to be in-kernel? or can it be module? Apr 04 17:47:33 just stuff required for botting Apr 04 17:47:38 *booting Apr 04 17:48:27 philipp64: usually, it's a module Apr 04 17:50:42 ok... the harder part is figuring what of the default stuff I now need to suppress... Apr 04 17:51:17 what kernel version do you currently build against? Apr 04 17:51:35 <{Nico}> check target/linux/x86/Makefile Apr 04 17:52:34 ok... the MFGPT and CS5535 stuff only appeared in 2.6.27, so we need that at a minimum... Apr 04 17:52:45 2.6.32.10 Apr 04 17:52:58 or .9 not sure Apr 04 17:53:03 .10 Apr 04 17:53:56 what is all the CONFIG_CPU_SUP_xxx stuff for? and what do I need to turn off? Apr 04 17:54:16 that must be new. Apr 04 17:54:52 and you turn off CONFIG_X86_BIOS_REBOOT? Apr 04 17:55:47 philipp64: it's support for different cpus iirc Apr 04 17:55:56 philipp64: and you can turn it off, if you don't have a standard bios Apr 04 17:57:07 what's CONFG_MTD? Apr 04 17:57:13 err... CONFIG_MTD? Apr 04 17:57:46 philipp64: you need it for flash Apr 04 17:58:03 memory technology device Apr 04 17:58:11 we were building without it in 2.6.27... Apr 04 17:59:23 it is essential for squashfs and jffs2 support Apr 04 17:59:29 ok. Apr 04 17:59:59 we support those filesystems through block2mtd emulation Apr 04 18:01:48 what I have is: http://www.pastebin.ca/1856753 Apr 04 18:01:59 seems incomplete. Apr 04 18:02:25 I guess I need to see what the generated config file looks like, and then diff them side-by-side from what we use in astlinux. Apr 04 18:03:08 is it enough to have "# CONFIG_ACPI is not set", or do I have to turn off all of CONFIG_ACPI_xxx as well? Apr 04 18:03:30 no, disabling the parent symbol should be enough Apr 04 18:03:38 philipp64: simply use an empty config file and build the kernel Apr 04 18:03:47 philipp64: it will ask for missing symbols Apr 04 18:03:53 yes Apr 04 18:04:11 no, I mean I want to see what the derived defaults with my deltas results in... Apr 04 18:04:20 make kernel_oldconfig helps after selecting x86 and your new subtarget in make menuconfig Apr 04 18:04:30 <{Nico}> not empty, keep CONFIG_MGEODE_LX=y Apr 04 18:06:13 {Nico}: eh? you lost me. I already have it. Apr 04 18:06:40 ok, so what goes in DEFAULT_PACKAGES ? Apr 04 18:08:03 philipp64: kmods required for your target Apr 04 18:08:05 do I also need (for instance) kmod-cs5535-gpio? or is that already set via config-default? Apr 04 18:08:55 nope, config-defaults basically defines which features the kernel image should contain and kmod-* are .ipk packages containing *.ko files Apr 04 18:08:57 <{Nico}> philipp64: i was following rtz & xMff suggestion Apr 04 18:13:55 no kmod-geodewdt ? Apr 04 18:16:00 seems like it is not packaged yet Apr 04 18:16:35 see package/kernel/modules/other.mk Apr 04 18:16:44 how is the *.ko called? Apr 04 18:16:52 and how the corresponding symbol? Apr 04 18:19:11 <{Nico}> xMff, philipp64 : it might better to put it in a ./target/linux/x86/modules.mk if it's x86 only Apr 04 18:20:02 not seeing that file. Apr 04 18:21:34 <{Nico}> philipp64: create it :) Apr 04 18:21:52 ok, where's the list of Submenu's? Apr 04 18:21:58 <{Nico}> same purpose as package/kernel/modules/*.mk, but target specific Apr 04 18:28:31 is there a WATCHDOG_MENU? Apr 04 18:29:16 <{Nico}> philipp64: feel free to create one, all those modules will need sorting at some point Apr 04 18:34:20 and the autoload order gets selected how? Apr 04 18:34:44 defined in the package definitions Apr 04 18:35:05 AUTOLOAD:=$(call AutoLoad,20,crc-ccitt) Apr 04 18:35:17 20 is the index Apr 04 18:35:27 right, but how do I know what the right ordering is? Apr 04 18:35:33 ah Apr 04 18:35:39 wekk Apr 04 18:35:45 *well, guess it Apr 04 18:35:57 10? Apr 04 18:36:16 the other watchdog drivers are loaded at 50 Apr 04 18:36:37 wow, that late? Apr 04 18:36:42 yeah Apr 04 18:37:20 does not matter actually since they arent enabled anyway until the userspace daemon activates them Apr 04 18:37:52 once userspace starts, the kmod loading should be complete already Apr 04 18:37:58 build #38 of adm5120 is complete: Failure [failed compile_6] Build details are at http://tksite.gotdns.org:8010/builders/adm5120/builds/38 Apr 04 18:39:52 make kernel_oldconfig failed because i486-openwrt-linux-uclibc-gcc didn't exist. Apr 04 18:40:03 missing dependency on toolchain? Apr 04 18:40:28 yes, you should have built the toolchain at least once Apr 04 18:41:26 what's the command? Apr 04 18:41:46 make toolchain/compile iirc Apr 04 18:43:49 ok, building toolchain... I have ftp://ftp.redfish-solutions.com/pub/net5501.patch Apr 04 18:45:29 <{Nico}> make toolchain/install Apr 04 18:45:48 philipp64: looks good so far, for consistency I'd suggest to rename geodewdt to geode-wdt Apr 04 18:46:31 philipp64: and the wdt package is not referenced in DEFAULT_PACKAGES Apr 04 18:46:39 nico * r20710 /packages/utils/collectd/Makefile: packages/collectd: fix typo in ascent configure hook, mark iptables & netlink plugins as broken on 2.4 (closes: #6961) Apr 04 18:47:00 philipp64: don't enable any modules in the config file Apr 04 18:47:12 philipp64: simply say not set Apr 04 18:48:35 ok, something rewrote that file! Apr 04 18:48:50 wasn't expecting that. Apr 04 18:49:50 philipp64: oldconfig does this Apr 04 18:51:31 can I temporarily override the LINUX_VERSION to make doing the side-by-side comparison easier> Apr 04 18:51:43 yes Apr 04 18:51:51 make ... LINUX_VERSION=... should work Apr 04 18:52:23 toolchain is still building. Apr 04 18:52:33 yeah, takes a while Apr 04 18:52:35 didn't know I needed quilt. Apr 04 18:52:51 openwrt uses quilt for patch management in various places Apr 04 18:52:55 btw, in astlinux, we had something to detect requisite tools at build-time. Apr 04 18:53:26 wouldn't the build break anyway? Apr 04 18:53:34 if they're not fulfilled I mean Apr 04 18:54:26 have a look at https://astlinux.svn.sourceforge.net/svnroot/astlinux/trunk/Makefile ... then look for "world: | host.flex", and "host.%:" Apr 04 18:55:48 and toolchain is still building. yow. takes 6 minutes for me on astlinux... but we build with -j5 Apr 04 18:57:06 xMff: the nice thing about the "world: | host.xxx host.yyy ..." rule is when you type "make" it checks for the presence of all required tools before it does anything else. Apr 04 18:57:30 hence, if anything is missing, it breaks right away... rather than getting partially into a build and then breaking. Apr 04 18:57:37 we have something similar, but without explicit distribution checks, since we also support building on non-linux stuff Apr 04 18:57:42 openwrt does the same, somebody just forgot to annotate the quilt requirement Apr 04 18:57:44 it tells you right away all the tools it's going to need up-front. Apr 04 18:58:08 quilt shouldn't be a prereq check, openwrt builds quilt Apr 04 18:58:10 we handle that too... Apr 04 18:58:19 look for BR2_HOST_DEBIAN ... Apr 04 18:58:33 nbd: not for me it didn't. Apr 04 18:58:51 but that might be because I didn't build toolchain first. Apr 04 18:59:01 it's part of tools/ Apr 04 18:59:21 and installed to staging_dir/host/ Apr 04 18:59:45 nbd: why does openwrt build it's own sed instead of using the system one? Apr 04 19:00:01 rtz2: there are many system versions of sed that are broken in interesting ways Apr 04 19:00:03 rtz2: cygwin braindeath? Apr 04 19:00:08 rtz2: because it could be buggy or incompatible, as every gnu util :P Apr 04 19:00:09 not just cygwin Apr 04 19:00:11 but bsd as well Apr 04 19:00:15 or darwin Apr 04 19:00:21 ahh, ok Apr 04 19:00:22 (which i build on) Apr 04 19:00:30 where's posix when we need it? Apr 04 19:00:35 not even "sort" sorts the same across arches :D Apr 04 19:00:49 philipp64: well, we use non-posix stuff, because strict posix often is too limited Apr 04 19:01:11 yeah, sort has some really ugly locale-dependent behaviour on linux Apr 04 19:02:22 philipp64: the usual entry point to compile is make world, this does the requried steps and stuff, could be that the prereq check was skipped because you started with a specific make target right away Apr 04 19:02:27 nbd: and another question Apr 04 19:02:43 nbd: did you see my mail with regard to the bcm4716? Apr 04 19:02:48 nbd: and the patch in it? Apr 04 19:04:13 i've seen it, but i didn't have time to review it yet. i'll check it once i'm back in germany again Apr 04 19:04:37 nbd: ok, thanks Apr 04 19:05:17 nbd: do you plan to submit the patch with the backplane reads to mainline? Apr 04 19:05:33 true. we had to do stuff like: $(LANG=C svn info | awk '/^Last Changed Rev:/ { print $4 }') Apr 04 19:05:54 and supporting building on debian with Dash was no fun... Apr 04 19:06:08 rtz2: which one specifically? Apr 04 19:06:37 philipp64: we just require bash, because there are many packages that depend on bash-isms Apr 04 19:06:57 the build process overrides SHELL Apr 04 19:07:28 rtz2: "make kernel_oldconfig" keeps blocking at GEODE_MFGPT_TIMER ... Apr 04 19:09:35 ok, now to build a kernel... Apr 04 19:09:54 posting new diffs... Apr 04 19:10:02 how do I build just the kernel? Apr 04 19:10:24 nbd: https://dev.openwrt.org/browser/trunk/target/linux/brcm47xx/patches-2.6.32/150-cpu_fixes.patch Apr 04 19:10:34 nbd: the stuff in r4kcache.h Apr 04 19:11:05 maybe Apr 04 19:11:11 right now it's still somewhat ugly Apr 04 19:11:13 I'm guessing "make target/install" Apr 04 19:11:22 yes Apr 04 19:11:32 nbd: it's also affected by this patch: https://dev.openwrt.org/browser/trunk/target/linux/brcm47xx/patches-2.6.32/920-cache-wround.patch Apr 04 19:11:45 forgot to override LINUX_VERSION and -j Apr 04 19:11:54 rtz2: yeah, those two should be merged Apr 04 19:12:51 nbd: I tried but couldn't get quilt to cooperate :) Apr 04 19:12:59 nbd: my patch also affects those two Apr 04 19:13:23 hmmmm "make kernel_oldconfig LINUX_VERSION=xxx" gets ignored. Apr 04 19:14:35 using quilt fold this should be easy Apr 04 19:15:17 nbd: I'm not exactly an expert for it Apr 04 19:15:26 ok, I have to eat now Apr 04 19:15:56 what's causing: Makefile:19: *** The generic kernel config for your kernel version is missing. Stop. ? Apr 04 19:16:04 if you are at platform/150-cpu_fixes.patch, run quilt fold < patches/platform/920-cache-wround.patch; quilt delete platform/920-cache-wround.patch; quilt refresh; ... Apr 04 19:16:18 philipp64: which kernel version are you using? Apr 04 19:16:27 2.6.26.42 Apr 04 19:16:35 oh, we don't support this kernel version anymore Apr 04 19:16:44 so there's no config, no patches, etc. in target/linux/generic-2.6/ Apr 04 19:17:08 why do you want to build this old version anyway? Apr 04 19:17:21 yow. I see that. you skip from 2.6.25 to 2.6.30! Apr 04 19:17:39 2.6.25 is only there because there's still a target using it Apr 04 19:17:44 we had .26 and other versions there Apr 04 19:17:47 (because the .config we use for the net5501 on astlinux is based on 2.6.27.42, and I want to diff against that.) Apr 04 19:17:49 but we removed them as soon as they became unused Apr 04 19:18:33 how about forward porting the astlinux config to .30 or .32 and then diffing against that? Apr 04 19:19:14 can't. I mentioned that yesterday on #openwrt... we use uclibc-0.9.28 so we're blocked on 2.6.27 until we do a uclibc version bump. Apr 04 19:19:30 ouch Apr 04 19:20:18 say, where do I set the -march? Apr 04 19:21:58 make menuconfig allows you to set it Apr 04 19:22:14 the defaults can be set in target/linux/x86/Makefile using the CFLAGS variable Apr 04 19:22:25 check other target makefiles if you want an example Apr 04 19:26:34 you don't use BR2_TARGET_OPTIMIZATION then... Apr 04 19:27:23 we don't use any BR2_* stuff Apr 04 19:27:40 OPTIMIZATION? Apr 04 19:27:53 CONFIG_TARGET_OPTIMIZATION Apr 04 19:28:03 which is filled in from CONFIG_DEFAULT_TARGET_OPTIMIZATION Apr 04 19:28:13 which is pulled from the CFLAGS in the target makefile by default Apr 04 19:41:41 cp: cannot stat `/home/philipp/openwrt/staging_dir/host/usr/lib/grub/i386-*/stage1': No such file or directory Apr 04 19:41:42 nbd: I will give it a shot Apr 04 19:42:21 <{Nico}> philipp64: make 'package/grub/compile' before 'make target/linux/install' Apr 04 20:03:43 another missing dependency? Apr 04 20:07:25 cp: cannot create regular file `/home/philipp/openwrt/bin/x86/openwrt-x86-vmlinuz': No such file or directory Apr 04 20:07:47 maybe you should just run make Apr 04 20:07:50 without arguments Apr 04 20:07:55 well except for kernel version Apr 04 20:14:52 nbd: does -j5 work? Apr 04 20:15:28 should Apr 04 20:50:05 ok, built. now how to get it into the box? Apr 04 20:50:46 philipp64: check image/makefile Apr 04 20:51:14 philipp64: best solution would be to create an image file for your box, which you can easily flash Apr 04 20:51:27 philipp64: if you device supports such a thing Apr 04 20:51:55 it's a boot-from-flash (CF) BIOS. Apr 04 20:52:16 philipp64: what's your flash layout? Apr 04 20:52:17 same as a net4801. Apr 04 20:52:26 it's not fixed. Apr 04 20:52:40 I can set it up however I like. Apr 04 20:53:46 philipp64: is the bios also in flash? Apr 04 20:54:05 no, it has it's own memory. Apr 04 20:54:30 philipp64: where does it expect the kernel (or how does it find the kernel)? Apr 04 20:55:20 syslinux + grub or syslinux + runnix Apr 04 20:55:27 rtz2: ordinary boot from ide, means bios starts grub, grub starts kernel Apr 04 20:55:52 philipp64: you want to place openwrt on flash or ide? Apr 04 20:55:56 philipp64: gunzip bin/*/openwrt-*.image.gz and dd it to your cf Apr 04 20:56:11 single partition? Apr 04 20:56:25 philipp64: this image contains grub and partition table, it is a disk image Apr 04 20:56:32 what size? Apr 04 20:56:43 I guess it doesn't matter as long as it's smaller than the CF Apr 04 20:56:59 60 by default iirc, you can increase that in menuconfig or change it with cfdisk later Apr 04 20:57:03 60 MB Apr 04 20:57:07 or 40 Apr 04 20:57:10 don't recall Apr 04 20:57:25 damn that's small. Apr 04 20:57:45 well its giant compared to other targets ;) Apr 04 20:57:56 << 3MB is normal Apr 04 21:00:55 how do you fit asterisk into that? Apr 04 21:01:16 you don't, unless you have external storage available Apr 04 21:01:34 like usb pen drives or something Apr 04 21:02:09 actually i think in the past some people did manage to fit asterisk on a 4M flash device Apr 04 21:02:28 yeah, you can always strip stuff down and leave other components out etc. Apr 04 21:03:17 btw, where do you set the default console baud for a platform? Apr 04 21:03:29 via menuconfig, target image options Apr 04 21:03:53 the defaults are in config.in Apr 04 21:04:20 target/linux/x86/image/Config.in Apr 04 21:09:17 booting... Apr 04 21:10:08 usbhid: v2.6:USB HID core driver Apr 04 21:10:15 and... it seems to hang. Apr 04 21:12:25 so what's the name of my target derived from? TARGET_x86_Soekris_net5501 ? Apr 04 21:12:59 the $BOARDNAME value? Apr 04 21:27:32 xMff: still here? Apr 04 21:30:42 rtz2: ping Apr 04 21:31:54 philipp64: pong Apr 04 21:32:48 trying to figure out why my changes to image/Config.in aren't resulting in the correct baud rate. Apr 04 21:34:09 philipp64: did you check the kernel command line, if it has a console argument? Apr 04 21:34:52 I think $(X86_GRUB_BAUDRATE) isn't being set correctly. Apr 04 21:35:53 where does SERIAL_CONFIG get synthesized? Apr 04 21:36:09 no idea Apr 04 21:36:16 philipp64: not my area, sorry Apr 04 21:36:36 arget/linux/x86/image/Makefile Apr 04 21:40:55 xMff: posted new diffs... but my CONFIG_X86_GRUB_BAUDRATE is still wrong. Apr 04 21:41:21 even though my .config has CONFIG_TARGET_x86_net5501=y in it. Apr 04 21:41:31 those defaults are only apllied if they have not been set yet Apr 04 21:41:52 if there is already some value in your toplevel .config, it will not be updated Apr 04 21:41:53 so... what do I need to do then? Apr 04 21:42:11 remove the .config and make menuconfig again? Apr 04 21:42:15 just change it, or start with a fresh config Apr 04 21:42:18 correct Apr 04 21:45:12 ok... next step: how to figure out what it was trying to do next after loading the usbhid module... i.e. when it hanged Apr 04 21:50:44 good question Apr 04 21:52:23 philipp64: try to load the module manually in failsafe and see, what happens Apr 04 21:52:27 log on http://www.pastebin.ca/1856991 Apr 04 21:53:32 ok, I'm at the prompt... Apr 04 21:53:36 <{Nico}> philipp64: sure it's hanging Apr 04 21:53:37 <{Nico}> :) Apr 04 21:54:00 <{Nico}> \o/ Apr 04 21:54:13 if I'm going to modprobe the modules, where's the default list? Apr 04 21:54:36 <{Nico}> look at /etc/modules.d/ Apr 04 21:54:46 <{Nico}> they should be loaded already Apr 04 21:55:10 lsmod shows nothing (this is safemode) Apr 04 21:55:19 philipp64: that's ok Apr 04 21:55:19 or failsafe or whatever. Apr 04 21:55:31 philipp64: and you have only insmod, no modprobe Apr 04 21:55:58 ok, so what to start with? usb-hid? Apr 04 21:56:33 (I also seem to have a bunch of ep80579-xxx stuff... not sure why!) Apr 04 22:23:52 build #37 of kirkwood is complete: Failure [failed compile_7] Build details are at http://tksite.gotdns.org:8010/builders/kirkwood/builds/37 Apr 05 00:26:00 rtz2: or anyone else... is there a way to run the /etc/rc (or whatever) script with -x ? Apr 05 00:26:22 rtz2: ping Apr 05 00:26:32 thepeople: pong Apr 05 00:26:40 philipp64: no idea, sorry Apr 05 00:26:43 rtz2: your access has been granted :-D Apr 05 00:27:06 well, I rebuilt from scratch, and ran it... apparently it had some stale ep80759 drivers that were messing things up... Apr 05 00:27:13 now it gets to: cs5535_gpio: can't allocate I/O for GPIO Apr 05 00:28:51 philipp64: well, you will have to debug the driver code Apr 05 00:29:24 but the same driver works fine in astlinux. Apr 05 00:30:53 thepeople: let me try :) Apr 05 00:32:24 /o\ Apr 05 00:34:42 this is what a running astlinux has: http://www.pastebin.org/135657 Apr 05 00:35:36 lines 27-$ are what's needed to support the platform... with the exception of the dahdi stuff, which is for PCI telephony cards. Apr 05 00:36:03 oh, and stp, llc, and bridge which are for bridging. Apr 05 00:38:51 hmm, for some reason svn doesn't like the username option :/ Apr 05 00:40:43 <{Nico}> rtz2: use openwrt Apr 05 00:40:44 rtz2: sorry, use openwrt not yours Apr 05 00:57:56 rtz2 * r20711 /packages/utils/nano/Makefile: set maintainer for nano Apr 05 00:58:10 yay Apr 05 00:58:14 :D Apr 05 01:02:37 :-D Apr 05 01:03:06 you even put it in the right place so I don't need to yell :-P Apr 05 01:03:29 thepeople: I checked where it belongs :) Apr 05 01:03:53 good, the first couple of commits to it had to be fixed Apr 05 01:03:55 :-) Apr 05 01:04:23 rtz2: btw, https://dev.openwrt.org/wiki/CommitPolicy Apr 05 01:59:08 nico * r20712 /packages/net/openswan/patches/130-compat_net_dev_ops.patch: [packages] openswan: add a patch to fix build on 2.6.30 targets where CONFIG_COMPAT_NET_DEV_OPS is not set Apr 05 02:15:16 abyrne * r20713 /packages/ipv6/wide-dhcpv6/Makefile: [packages] wide-dhcpv6: Set maintainer Apr 05 02:33:03 who knows the startup procedure well? trying to figure out why my boot is hanging... Apr 05 02:34:31 <{Nico}> philipp64: try asking cshore, he contributed the new preinit/init code Apr 05 02:39:00 nico * r20714 /trunk/target/linux/rdc/config-2.6.30: target/rdc: revert [20653], the proper fix for #7005 is in [20712] Apr 05 02:40:11 holy bloody crap Apr 05 02:42:12 you guys like changed.... EVERYTHING since r14214 Apr 05 02:42:20 i need to update more often -_- Apr 05 02:42:58 Weedy: well, yea :-D Apr 05 02:44:29 yey for broken network on boot Apr 05 02:45:44 nico * r20715 /branches/backfire/target/linux/rdc/config-2.6.30: [backfire] target/rdc: backport r20714 Apr 05 02:47:31 o..O **** ENDING LOGGING AT Mon Apr 05 02:59:56 2010