**** BEGIN LOGGING AT Sun Sep 19 02:59:56 2010
Sep 19 03:04:15 <xxiao_> grub 0.97 fails on x86 using host gcc(ubuntu 10.04), need add CFLAGS=-fno-stack-protector, looks like ubuntu's gcc has that by default
Sep 19 03:17:47 <xxiao_> hmm i saw no-stack-protector is there for host-gcc, however grub still fails
Sep 19 03:18:03 <xxiao_> bios.c:(.text+0x1d5): undefined reference to `__stack_chk_fail'
Sep 19 03:18:24 <xxiao_> builtins.c:(.text+0x7ac): undefined reference to `__stack_chk_fail'
Sep 19 04:56:27 <philipp64|laptop> goodasGNU: sent an email... hopefully someone will see it and answer.
Sep 19 05:33:46 <CIA-14> acoul * r23085 /trunk/package/zlib/Makefile: package/zlib: fix Darwin compile failure (closes #7963)
Sep 19 11:18:32 <CIA-14> florian * r23086 /trunk/package/kernel/modules/netdevices.mk:
Sep 19 11:18:32 <CIA-14> [package] Fix missing dependency in solos-pci module
Sep 19 11:18:32 <CIA-14> Modinfo for solos-pci tells us that it has a dependency on "atm", but this isn't
Sep 19 11:18:32 <CIA-14> captured in the Makefile.
Sep 19 11:18:32 <CIA-14> Signed-off-by: Philip A. Prindeville <philipp_subx@redfish-solutions.com>
Sep 19 11:31:37 <builder> build #117 of brcm63xx is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/brcm63xx/builds/117
Sep 19 13:12:32 <xMff> ping stintel
Sep 19 13:14:28 <xMff> nvm, found it
Sep 19 13:39:10 <stintel> :)
Sep 19 13:52:12 <xMff> it was about your udhcpc kill patch
Sep 19 13:52:41 <stintel> ah :)
Sep 19 13:53:33 <xMff> I extended the kill part a bit, first term, then wait up to 5 seconds, then kill
Sep 19 13:54:10 <xMff> going to wrap that into a procedure and put it into /etc/functions.sh, the same code is used in at least three places now
Sep 19 13:54:30 <stintel> ok, sounds good
Sep 19 13:54:57 <xMff> and I discovered one more race :(
Sep 19 13:55:55 <stintel> oh well, good that you discovered it ;-)
Sep 19 13:56:40 <stintel> not that bad as for example CVE-2010-3301
Sep 19 13:56:51 <stintel> I was quite shocked when I read about it
Sep 19 13:57:04 <xMff> its fun... every time I get a faster toy I discover new races in init
Sep 19 13:57:23 <xMff> yeah the cve is bad
Sep 19 13:57:38 <xMff> fortunately we have no 64bit targets in openwrt so no need to take any action
Sep 19 13:58:08 <xMff> interesting though that this is the thrid time in a row or so where security fixes got reverted
Sep 19 13:58:22 <xMff> I guess they don't commented code well enough
Sep 19 13:58:28 <xMff> didn't comment
Sep 19 13:59:06 <stintel> tried the exploit, I believe it works on all my systems except my centos 5 ones
Sep 19 13:59:19 <stintel> all my 64bit systems
Sep 19 13:59:20 <stintel> *
Sep 19 14:00:23 <xMff> well... shit happens
Sep 19 14:00:41 <stintel> luckily the few users who have accounts on one of my boxes don't know anything about security stuff
Sep 19 14:00:52 <stintel> yeah, shit happens, this was a long one :P
Sep 19 14:01:16 <xMff> ;)
Sep 19 14:04:04 <CIA-14> jow * r23087 /trunk/package/base-files/ (Makefile files/etc/functions.sh): [package] base-files: introduce service_kill() into functions.sh - a convenience wrapper for killing services by pid or pidfile with a grace period for termination before kill
Sep 19 14:06:18 <stintel> good timing, I'll need to use my rspro in client mode soon, cancelled my vdsl2 connection, end date is 21/09 and I didn't get to ordering a cable connection yet :/
Sep 19 14:09:57 <CIA-14> jow * r23088 /trunk/package/ppp/ (Makefile files/ppp.sh): [package] ppp: use service_kill()
Sep 19 14:15:54 <CIA-14> swalker * r23089 /packages/net/ (batman/Makefile bmxd/Makefile): [packages] batman, bmxd: change urls for the open-mesh.net -> open-mesh.org move
Sep 19 14:21:40 <xMff> stintel: hm I believe the removed locking for udhcpc is what introduced the race in the first place
Sep 19 14:22:00 <xMff> right now I see iptables: resource temporarely unavailable if the dhcp server answers really fast
Sep 19 14:23:54 <stintel> you mean with my patch applied ?
Sep 19 14:24:00 <xMff> yes
Sep 19 14:24:16 <stintel> ah, sorry 'bout that
Sep 19 14:24:34 <xMff> in the particular case of udhcpc it can happen that ifdown still runs while the new instance does ifup
Sep 19 14:27:53 <xMff> hm no
Sep 19 14:28:10 <xMff> its not stop vs start but lan vs wan
Sep 19 14:37:23 <xMff> thats odd
Sep 19 14:37:37 <xMff> I cannot reproduce with ifup -a; wifi while with /e/i/network restart I can
Sep 19 14:37:43 <xMff> they're basically doing the same
Sep 19 14:38:58 <stintel> xMff: you mean it occurs if you use dhcp on both lan and wan?
Sep 19 14:39:09 <xMff> no default
Sep 19 14:39:12 <xMff> lan static wan dhcp
Sep 19 14:48:30 <stintel> hmm, I always just ran "wifi" so maybe that explains why I didn't run into it while testing my changes
Sep 19 14:48:42 <stintel> and can't test right now, my 2nd rspro is at work
Sep 19 14:50:32 <stintel> brb, hungry
Sep 19 14:50:42 <swalker> updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/uscan.shtml
Sep 19 15:02:18 <CIA-14> jow * r23090 /trunk/package/firewall/ (Makefile files/lib/core_interface.sh): [package] firewall: protect iptables invocations with locks in interface ops, it might run concurrently due to hotplug invocations on network restart
Sep 19 15:04:59 <CIA-14> jow * r23091 /trunk/package/base-files/files/usr/share/udhcpc/default.script:
Sep 19 15:04:59 <CIA-14> [package] base-files: cached state vars after scan_interfaces might be out of
Sep 19 15:04:59 <CIA-14> sync when binding dhcp leases, this could lead to an unconfigured wan interface
Sep 19 15:04:59 <CIA-14> in certain circumstances. Replace config_get calls with uci_get_state
Sep 19 15:13:36 <CIA-14> jow * r23092 /trunk/package/base-files/files/lib/network/config.sh:
Sep 19 15:13:36 <CIA-14> [package] base-files: kill possibly existing udhcpc instance when bringing up a dhcp iface, utilize service_kill() in dhcp shutdown.
Sep 19 15:13:36 <CIA-14> Based on patch by Stijn Tintel. Possibly related to #7314
Sep 19 15:14:05 <stintel> seems you found it :)
Sep 19 15:15:26 <CIA-14> jow * r23093 /trunk/package/base-files/Makefile: [package] base-files: bump pkg revision after r23091, r23092
Sep 19 15:16:53 <builder> build #93 of adm5120 is complete: Failure [failed shell_3]  Build details are at http://tksite.gotdns.org:8010/builders/adm5120/builds/93
Sep 19 15:18:38 <CIA-14> jow * r23094 /branches/backfire/package/base-files/ (4 files in 4 dirs): [backfire] merge r23087, r23091 and r23092
Sep 19 15:19:23 <CIA-14> jow * r23095 /branches/backfire/package/ppp/ (Makefile files/ppp.sh): [backfire] merge r23088
Sep 19 15:21:28 <xMff> cshore: the preinit merge stuff is ready so far
Sep 19 15:22:07 <xMff> cshore: someone needs to rewrite block-extroot
Sep 19 15:27:27 <cshore> xMff: ok, I see if I can find some time...can you post the merge stuff you've done so far?
Sep 19 15:28:36 <xMff> http://luci.subsignal.org/~jow/preinit-merge-overlay.patch
Sep 19 15:29:58 <cshore> I'm busy for a couple of hours, but I'll see what I can do in later today
Sep 19 15:30:16 <xMff> the main culprit I've had so far is that if preinit dies, the kernel panics
Sep 19 15:30:55 <cshore> yeah, I don't know what to do about that...syntax errors have to be detected before flashing
Sep 19 15:31:03 <xMff> assuming that the end of /etc/preinit is never reached if /sbin/init got executed sucessfully you might consider adding a telnetd invocation or similar there
Sep 19 15:31:36 <xMff> hm syntax errors yes but not runtime errors
Sep 19 15:31:41 <cshore> would that be true for initramfs?
Sep 19 15:31:58 <xMff> don't know right now
Sep 19 15:32:04 <cshore> have you had runtime errors causing preinit to die?
Sep 19 15:32:22 <xMff> yes as soon as I enabled block-extroot
Sep 19 15:32:36 <xMff> while adding echos and other stuff to existing scripts worked as expected
Sep 19 15:33:33 <xMff>     . /lib/functions/extmount.sh
Sep 19 15:33:33 <xMff>     . /lib/functions/mount.sh
Sep 19 15:33:46 <xMff> thats what makes it fail I believe
Sep 19 15:34:02 <cshore> that makes sense....it's absolute
Sep 19 15:34:14 <cshore> but it's on the overlay now
Sep 19 15:35:14 <cshore> and it happens at the sourcing of preinit time I believe...errors when sourcing preinit causes preinit to fail with a kernel panic
Sep 19 15:35:22 <xMff> correct
Sep 19 15:35:28 <xMff> no
Sep 19 15:35:34 <xMff> it happens during execution
Sep 19 15:35:47 <xMff> I basically have a state with working ping but no telnet, sshd or httpd
Sep 19 15:35:55 <xMff> so preinit partly ran
Sep 19 15:36:05 <cshore> right...oh so not kernel panic?
Sep 19 15:36:35 <xMff> not sure anymore, maybe I mix stuff together here
Sep 19 15:36:48 <xMff> it should panic
Sep 19 15:38:31 <cshore> I would have thought only syntax errors would cause a panic, while other errors cause failure of some other kind
Sep 19 15:38:57 <cshore> our does sourcing failure cause the entire script to fail?
Sep 19 15:41:51 <xMff> yep
Sep 19 15:42:33 <xMff> http://openwrt.pastebin.com/Rg7THxYV
Sep 19 15:51:26 <CIA-14> jow * r23096 /packages/net/miniupnpd/ (Makefile files/miniupnpd.hotplug): [package] miniupnpd: fix iptables rule order (#7965)
Sep 19 15:55:11 <xMff> cshore: you should do a pi_include which at least warns if the file does not exist
Sep 19 15:55:33 <xMff> ideally it would first check /tmp/overlay/ then /
Sep 19 15:58:27 <cshore> hmmm....maybe we should set the /tmp/overlay in a variable that is set in the file that mounts the jffs
Sep 19 15:58:42 <cshore> xMff: so subsequent functions know they are on the overlay
Sep 19 16:04:31 <xMff> cshore: it should always do both
Sep 19 16:04:40 <xMff> sometimes functions are not present in the overlay
Sep 19 17:23:46 <CIA-14> nbd * r23097 /trunk/package/mac80211/patches/ (521-ath9k_aggr_race_fix.patch 522-ath9k_aggr_flush.patch): ath9k: fix various aggregation related race conditions
Sep 19 18:45:32 <builder> build #106 of octeon is complete: Failure [failed compile_7]  Build details are at http://tksite.gotdns.org:8010/builders/octeon/builds/106
Sep 19 19:38:30 <dtatulea_> hi. does the linux kernel archive contain the patches? if not, where can i find them?
Sep 19 19:39:26 <xMff> what patches?
Sep 19 19:40:31 <dtatulea_> isn't the kernel patched?
Sep 19 19:41:59 <xMff> you'll find the openwrt kernel patches in target/linux/generic/patches-* and target/linux/arch/patches-*
Sep 19 19:48:32 <dtatulea_> xMff: thanks!
Sep 19 19:52:14 <cshore> ok, I'll be around for working on extroot in little bit
Sep 19 19:52:27 <xMff> \o/
Sep 19 19:58:42 <cshore> xMff: sorry I just had a change of plans...I'll be another hour or so...my sister wants to go for a walk in the park with me and my dog before she heads home
Sep 19 19:59:00 <xMff> k
Sep 19 21:13:33 <nunojpg> when package A and B are defined in the same Makefile, and package A is selected, package B dependencies are also built?
Sep 19 21:14:07 <nunojpg> I'm seeing them being compiled(but not installed), and I don't see any reason for that
Sep 19 21:26:39 <OPTY_laptop> xMff, got a minute?
Sep 19 21:27:09 <OPTY_laptop> how does the program diff work?
Sep 19 21:27:19 <OPTY_laptop> is it 'diff OLD NEW'
Sep 19 21:27:32 <xMff> yes
Sep 19 21:27:32 <OPTY_laptop> or 'diff new old' ?
Sep 19 21:27:37 <xMff> old new
Sep 19 21:27:45 <OPTY_laptop> kk ty
Sep 19 21:29:19 <OPTY_laptop> i cant believe that xchat changes ppl to use the app on windows LOL
Sep 19 21:29:29 <OPTY_laptop> charges*
Sep 19 21:36:32 <builder> build #88 of rb532 is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/rb532/builds/88
Sep 19 21:52:10 <FrozenSpoon> I'm trying to add a new package (locally) with patches. I created a patches/ directory and put .patch files in it. The build system doesn't seem to apply these patches automatically. Other packages don't seem to have special magic defined for patches and they happen automatically, am I missing something?
Sep 19 21:52:36 <xMff> patches appliable with -p1 ?
Sep 19 21:53:04 <FrozenSpoon> hmmm probably not, but even with V=99 I wasn't seeing it even attempt to patch
Sep 19 21:53:32 <xMff> there's no special magic for patches
Sep 19 21:59:05 <FrozenSpoon> but it should attempt to apply them automatically, right? In other packages, I don't see anything special in their Makefile to make it happen
Sep 19 21:59:17 <xMff> yes
Sep 19 22:07:38 <TheRealOpty> hmm wtf
Sep 19 22:09:58 <cshore> xMff: ok, I'm here
Sep 19 22:10:33 <xMff> long walk :)  must've been dark by now
Sep 19 22:10:56 <cshore> it's only 6pm here
Sep 19 22:11:02 <xMff> oh, right
Sep 19 22:11:04 <xMff> nvm
Sep 19 22:13:19 <cshore> xMff: do you want extroot to be moved to base system?
Sep 19 22:13:26 <xMff> yes
Sep 19 22:13:54 <xMff> its not really a "disc util"
Sep 19 22:14:33 <cshore> I didn't know where else to put it, and at the time it wasn't a core thing
Sep 19 22:14:35 <xMff> svn up
Sep 19 22:14:37 <xMff> oops
Sep 19 22:14:54 <cshore> do you want to move extroot or shall I?
Sep 19 22:15:06 <xMff> you can do it
Sep 19 22:15:11 <FrozenSpoon> well I dunno, I must simply be missing something that triggers the magic http://pastebin.com/gcC7c6a9
Sep 19 22:15:51 <xMff> FrozenSpoon: it has no external source
Sep 19 22:15:58 <xMff> therfore no patches are applied
Sep 19 22:16:27 <xMff> actually it is because you redefined Build/Prepare
Sep 19 22:16:38 <xMff> which is implied by internal src
Sep 19 22:18:14 <FrozenSpoon> ah yeah, if I get rid of my Build/Prepare it does apply it
Sep 19 22:23:18 <FrozenSpoon> I take that back, actually
Sep 19 22:23:57 <cshore> xMff: oh, what happens in terms of Package/<package>/install for the ImageBuilder?  I think it may be an issue for the creates of /lib/preinit/00_extroot.conf, and for regular preinit for /lib/preinit/00_preinit.con in base-files
Sep 19 22:24:15 <xMff> you mena postinstall?
Sep 19 22:24:25 <cshore> no, the package build
Sep 19 22:24:34 <cshore> the install into .ipkg
Sep 19 22:25:13 <xMff> imagebuilder works with precompiled ipkg files
Sep 19 22:25:25 <cshore> oh, ok, so it should be okay
Sep 19 22:29:33 <xMff> heh wondering why extroot does not work
Sep 19 22:29:43 <xMff> guess I should insert the thumbdrive
Sep 19 22:29:49 <cshore> heh
Sep 19 22:30:47 <cshore> the fallback works too well?
Sep 19 22:32:35 <xMff> yeah
Sep 19 22:34:38 <CIA-14> cshore * r23098 /trunk/package/block-extroot/Makefile: [package] base/block-extroot: Move block-extroot in menuconfig from Utilities/disc to Base system
Sep 19 22:35:29 <TheRealOpty> cshore, update wiki ^
Sep 19 22:35:35 <TheRealOpty> :)
Sep 19 22:35:54 <cshore> TheRealOpty: do you mind doing it?
Sep 19 22:36:22 <TheRealOpty> xMff, that bug isnt fixed. the one regarding fdisk, 'no-input-files'
Sep 19 22:36:35 <TheRealOpty> cshore, sure ill get to it here soon
Sep 19 22:37:36 <FrozenSpoon> okay thanks for the help xMff, I think I understand now
Sep 19 22:39:44 <FrozenSpoon> btw guys this openwrt thing is pretty great, was really amazed when the first image I built on trunk installed and ran perfectly
Sep 19 22:40:44 <xMff> TheRealOpty: it is. you need to refresh your .config
Sep 19 22:43:00 <RealOpty> xMff, for libncurses to be added yeah?
Sep 19 22:43:41 <RealOpty> I was doing a fresh trunk for x86 and reproduced that bug again.
Sep 19 22:44:05 <RealOpty> philipp64|laptop reproduced the bug too
Sep 19 22:44:23 <philipp64|laptop> yup. guilty.
Sep 19 23:16:29 <cshore> xMff: any objection to my adding http://openwrt.pastebin.ca/1944595 to /lib/functions/boot.sh in base-files?
Sep 19 23:19:26 <xMff> cshore: nope, thats fine
Sep 19 23:19:51 <xMff> I'd replace -e with -f though
Sep 19 23:20:11 <xMff> to rule out attempts to source a directory
Sep 19 23:20:24 <cshore> right...hadn't thought of that
Sep 19 23:21:05 <xMff> the srcfile assignment is not needed
Sep 19 23:21:20 <xMff> and change line 4 to quote the whole string
Sep 19 23:21:27 <xMff> looks better...
Sep 19 23:22:36 <cshore> http://openwrt.pastebin.ca/1944605
Sep 19 23:25:22 <xMff> yeag
Sep 19 23:25:28 <xMff> srcfile is still redundant
Sep 19 23:25:40 <cshore> doh
Sep 19 23:26:14 <cshore> http://openwrt.pastebin.ca/1944606
Sep 19 23:26:20 <cshore> or should I use srcfile?
Sep 19 23:26:36 <xMff> no its okay
Sep 19 23:30:15 <CIA-14> cshore * r23099 /trunk/package/base-files/files/lib/functions/boot.sh:
Sep 19 23:30:15 <CIA-14> [package] base-files: Add function pi_include the /lib/functions/boot.sh used by
Sep 19 23:30:15 <CIA-14> preinit, which safely sources files as well as working correctly whether the
Sep 19 23:30:15 <CIA-14> script using it is a package added to the jffs2 part of a squashfs or built into
Sep 19 23:30:15 <CIA-14> an image or on some other root filesystem.
Sep 20 00:04:06 <RealOpty> cshore, so that ^ makes it so one can install and use  extroot after the image is built?
Sep 20 00:04:37 <cshore> RealOpty: not but it's a prerequisite function
Sep 20 00:04:43 <cshore> *no
Sep 20 00:04:59 <RealOpty> ah ok
Sep 20 00:05:11 <RealOpty> cshore, but are you working on making that possible?
Sep 20 00:05:18 <cshore> I'm off to bed....I work on this when I wake up
Sep 20 00:05:20 <cshore> yes
Sep 20 00:05:28 <RealOpty> coolbeans
Sep 20 00:05:30 <RealOpty> night
Sep 20 01:10:38 <philipp64|laptop> trying to build trunk with eglibc, but it's puking. /home/philipp/openwrt/build_dir/toolchain-i386_gcc-4.1.2_eglibc-2.8/eglibc-2.8-r8587-initial/config.h:3:3: error: #error "glibc cannot be compiled without optimization"
Sep 20 01:10:44 <philipp64|laptop> any ideas what I need to do?
Sep 20 02:16:51 <philipp64|laptop> [florian]: ping
Sep 20 02:29:51 <RealOpty> xMff, should r23098 be merged into backfire?
**** ENDING LOGGING AT Mon Sep 20 02:59:57 2010