[00:24] <Barhine> whoah
[00:24] <Barhine> this is a chennel
[00:24] <Barhine> so how far along is ubuntu on arm?
[05:04] <BitWraith> does anybody know where I can find a version of kexecboot that supports ext4? or am I going to have to install all over again because I used such a new filesystem? :-/
[05:55] <dpb> That's the reason why people have usually made a separate /boot partitions with an older filesystem.
[07:20] <BitWraith> dpb, you ca't do that with kexecboot, it passes its own root= option to he kernel forcing it to use the same partition the kernel image was found on
[07:22] <BitWraith> supposedly they are fixing that in the latest version, but there aren't any ready-to-flash images of it yet, and I haven't figured out how t compile it myself yet
[07:24] <dpb> I see.
[11:19] <lool> ogra: Hmm we should be here
[11:19] <ogra> heh, yeah
[11:21] <ogra> hmm, identical version in debian
[11:21] <ogra> thats intresting
[11:43] <lool> ogra: What about the PR 10288 change?
[11:44] <lool> ogra: I couldn't fetch anything newer than 3rd of July as CVS seems down for m
[11:44] <lool> me
[11:44] <ogra> i'm just building with that one disabled :)
[11:44] <lool> Are you bisecting?
[11:44] <ogra> no, i just commented out http://paste.ubuntu.com/216821/
[11:44] <ogra> s/commented out/reverted
[11:45] <ogra> its at the ld tests ... i'll know in 5 mins
[11:45] <lool> I doubt it's that change; I'd rather suspect the testsuite
[11:45] <ogra> not being brought up to date with the code you mean ?
[11:46] <ogra> well, ten it should still survive this test now
[11:46] <ogra> *then
[11:46] <lool> The testsuite was completely rewritten
[11:47] <ogra> between 20090704 and 20090622 ?
[11:47] <lool> Yes
[11:49] <ogra> i only see two changelog entries
[11:49] <ogra> +	* binutils-all/objcopy.exp: Move XFAIL from objcopy_test to
[11:49] <ogra> +	copy_executable.
[11:49] <ogra> and another unrealted one pointing to cygwin
[11:49] <lool> There are multiple ChangeLog file
[11:50] <lool> ogra: Did you file a bug already?
[11:50] <ogra> yeah, i just see that
[11:50] <ogra> nope
[11:50] <lool> ogra: You might want to start there
[11:50] <ogra> i was hoping to be able to attach some info
[11:50] <lool> Start by filing the bug, then collect more info?
[11:51] <ogra> well, the build is near the point of failure ... i'll file after it failed or passed that
[11:59] <ogra> ok, failed
[11:59]  * ogra files ...
[12:02] <ogra> oh
[12:02]  * ogra looks at http://sourceware.org/ml/binutils/2009-06/msg00289.html
[12:10] <ogra> lool, bug 398732
[12:10] <ubot4> Launchpad bug 398732 in binutils "binutils on armel fails to build due to regression test failure on 64 Bytes alignment test for gas/arm/arm.exp" [High,New] https://launchpad.net/bugs/398732
[12:11] <ogra> feel free to mark it critical if you think its worth it
[12:12] <lool> Thanks
[12:12] <lool> ogra: I was thinking of filing it upstream though
[12:13] <ogra> i want to have it at dokos attention first, i'll file an upstream one as well
[12:16] <ogra> sigh, indeed upstream bugzilla needs a new account
[12:16]  * ogra hates that
[12:40] <neure> hello
[12:40] <neure> anyone got kernel config for qemu?
[12:44] <ogra> the linux-image for versatile package contains the qemu config we use
[12:47] <ogra> lool, wow, upstream is quick :) already confirmed
[12:48] <lool> ogra: Where is that confirmed?
[12:48] <ogra> in the LP bugwatch
[12:50] <lool> That seems broken, I see no activity in the upstream bug
[12:51] <ogra> yeah, weird
[12:56] <ogra> lool, btw, are you using the packaged kernel in your cloud builder or a self rolled one (and did your chnages get included in our tree already) ?
[12:57] <lool> ogra: You mean the qemu kernel?
[12:57] <ogra> i'd like to switch rootstock to using the packaged version and new qemu
[12:57] <ogra> so i have a VM matching the HW we support
[12:57] <lool> I will probably use a custom kernel and qemu, but the Ubuntu ones are good enough for now
[12:58] <ogra> well, i dont want to have to change twice if we switch to v6
[12:58] <lool> ogra: We don't have a versatile flavour in karmic anymore
[12:59] <ogra> ugh
[12:59] <ogra> ok
[12:59] <lool> The problem is that the upstream linux doesn't support arm > v5 versatile by default
[12:59] <ogra> i thought that was carried over
[12:59] <lool> It was dropped along the "drop all the v5 stuff" commit
[12:59] <ogra> hrm
[12:59] <ogra> we could really have kept that
[13:00] <lool> I don't have time to fix the kernel build to allow both v5 and v6/v7; if you'd like to chase it, I think it's a good idea
[13:00] <ogra> given that we need it back anyway
[13:00] <ogra> lets see ... if we really skip A3 i migh have time for it
[13:01] <lool> ../../ld/ldlang.c:1399: error: no previous prototype for ‘next_matching_output_section_statement’
[13:01] <ogra> for now there are still binutils, banshee, tomboy and a bunch of specs i have to attack :)
[13:01] <ogra> not to forget the ftbfs list :)
[13:02] <neure> ogra, where is that image?
[13:02] <ogra> in the archive
[13:02] <ogra> ports.ubuntu.com
[13:04] <neure> sorry looks like i dont know how to use that
[13:06] <ogra> http://ports.ubuntu.com/pool/main/l/linux/linux-image-2.6.28-14-versatile_2.6.28-14.46_armel.deb
[13:10] <neure> erm
[13:10] <neure> package architecture (armel) does not match system (i386)
[13:11] <neure> i wonder how can i get that open just to the the config file?)
[13:11] <ogra> open it with file roller :)
[13:13] <neure> ah
[13:13] <neure> got config-2.6.28-14-versatile
[13:13] <neure> is that the one you use with qemu?
[13:13] <neure> can i simply use it with 2.6.30 as well?
[13:13] <ogra> you probably need to make changes
[13:13] <ogra> but its a good base
[13:15] <neure> im having trouble getting anything into the qemu console on boot
[13:27] <neure> well
[13:27] <neure> i managed to build the kernel
[13:27] <neure> and i get to login
[13:28] <neure> but my question is: why i dont see kernel boot messages?
[13:28] <ogra> you likely dont use the right console= bootoption
[13:29] <neure> well
[13:29] <neure> i've tried many
[13:30] <ogra> console=ttyAMA0,115200n8 should be the right one looking at http://bazaar.launchpad.net/~ogra/project-rootstock/trunk/annotate/head%3A/rootstock
[13:31] <neure> ive tried console=ttyAMA0
[13:31] <neure> i wonder if ,115200n8 is important, but I can try
[13:31] <neure> first im enabling preempting in the kernel..
[13:32] <ogra> yes, it is
[13:32] <neure> nope, that didnt help
[13:49] <neure> oh wait
[13:49] <neure> now i do gte the boot messages if i switch to ctrl-alt-3
[13:49] <neure> i wonder - can I get that to the default, ctrl alt 1 console?
[13:52] <lool> Rebuilding latest binutils under jaunty gives a passing testsuite, so it builds with gcc 4.3
[13:53] <ogra> lool, well, given it finishes fine in debian i wouldnt have doubted that
[13:55] <ogra> fsvo fine :) there are tons of error messages from gcc in their log
[14:42] <lool> ogra: http://sourceware.org/ml/binutils/2009-07/msg00038.html
[14:42] <lool> (thanks to doko)
[14:42] <ogra> yeah
[14:44] <ogra> i dont see the patch anywhere though
[14:45] <ogra> ah
[14:45] <ogra> found it
[14:46] <ogra> http://sourceware.org/ml/binutils/2009-06/msg00289.html
[14:46] <lool> That's the one you mentionned earlier
[14:46] <lool> Just Matthias reported the issue already
[14:46] <ogra> yep
[14:46] <ogra> ok, but didnt open any bugs
[15:08] <lool> For some reason I can't build that very well from tip
[15:17] <lool> Hey doko
[15:19] <lool> I reproduce the failure after adding the patch
[15:19] <lool> doko: So you spotted the right patch, thanks1
[15:19] <lool> !
[15:24] <doko> lool: yes, but Daniel Gutson wasn't able to reproduce it
[15:38] <ogra> doko, do you think its not gcc version related at all ? i dont see him building with 4.4
[15:38]  * ogra had the impression its the combo of patch and gcc version
[15:38] <lool> doko: Could you bounce me his message so that I can reply to it not breaking the thread?
[15:38] <lool> doko: I'm sub-ing to the list
[15:39] <doko> lool: how do you bounce from thunderbird?
[15:40]  * ogra thinks TB needs a special plugin
[15:41] <ogra> mail redirect plugin it seems
[15:45] <lool> doko: I don't know :)
[15:45] <lool> doko: forward as an attach?
[15:54] <lool> doko: So shall we just push a binutils without that patch?
[15:54] <lool> doko: I confirmed that this single patch is causing the regression by building tip + that patch
[15:54] <doko> lool: which regression exactly do you mean?
[15:55] <lool> doko: The one from the testsuite
[15:55] <doko> the newly added test?
[15:55] <lool> Yes
[15:55] <lool> Or do you think we should fix the test ourselves?
[15:55] <ogra> would be good to have a working ld :)
[15:55] <ogra> one way or the other
[15:55] <doko> well, that patch is a patch for as, so why is ld broken?
[15:56] <lool> doko: I'm just fixing the binutils build so that we have latest binutils
[15:56] <ogra> probably the ld issue i see in the upstart build is just a fallout
[15:56] <doko> lool: please let my build finish
[15:57] <lool> doko: Sure, so you're building a binutils without that patch?
[15:59] <doko> I'm currently checking current trunk with this patch, and trying to find out why it's not seen by Daniel. plus building on Debian as well
[16:00] <doko> and we did turn on ssp recently on arm as well ...
[16:20] <doko> lool, ogra: please could you check if you see the ld failure with the 20090620 binutils? This one doesn't fail the test
[16:22] <doko> now building the debian binutils with the align patch on unstable. it's a slow machine ...
[16:22] <ogra> will doo, but it will take a while, SD filesystem on my build machine here
[16:30] <lool> doko: Isn't that the one we have now?  we have 20090622
[16:30] <ogra> doko, 20090620 ? 20090622 doesnt fail it either i think
[16:30] <ogra> heh, snap
[16:31] <lool> doko: I see you did upload a 20090620; I guess I could try that
[16:31] <ogra> though 22 did build fine ...
[16:31] <ogra> (but breaks in ld for the upstart build)
[16:32] <ogra> so 22 either didnt have the test or passed it
[16:42] <doko> lool: got it, changed objdump output, which breaks the comparision with the expected output
[16:47] <lool> doko: aha
[16:48] <lool> doko: But I only see it with the patch
[16:54] <doko> lool: the testcase is added with the patch
[17:01] <lool> doko: So I guess we should close the upstream bug and send an updated patch to Daniel Gutson?
[17:02] <lool> doko: I need to reboot my B2; the upstart build hung it; I guess OOM
[17:04] <doko> ouch, ok, not sure if I had something building on it
[17:05] <ogra> lool, your B2 ?
[17:05] <ogra> where did you get a B2 from ?
[17:05]  * ogra thought we only had one left
[17:05] <lool> ogra: Just like you did?!
[17:05] <ogra> oh, you didnt send yours to the kernel team
[17:05]  * ogra forgot
[17:06] <lool>  3954 lool      20   0  135m 126m 6216 R 92.9 26.8   0:38.71 cc1
[17:06] <ogra> you were hiding for so long on that spanish island that i actually forgot about yours :)
[17:06] <ogra> it hangs in cc ? not in ld ?
[17:06] <ogra> wow, thats weird
[17:09] <lool> 16750 nih-dbus-tool/tests/test_node.c
[17:10] <lool> It's just a stupidly large file
[17:10] <ogra> it builds in less than 5min on x86
[17:10] <ogra> the whole of upstart
[17:10] <lool> But it might be using a huge amount of memory
[17:11] <ogra> hmm, i didnt check that
[17:11] <ogra> i think Keybuk said something about 16000 LOC
[17:18] <NCommander> o___________O;
[17:34] <lool> ogra: Did you sub ubuntu-armel on that upstart bug?
[17:34]  * ogra thought he did
[17:34] <ogra> https://bugs.launchpad.net/ubuntu/+source/gcc-4.4/+bug/398403
[17:34] <ogra> yep
[17:34] <ubot4> Launchpad bug 398403 in gcc-4.4 "gcc-4.4 fails to build upstart 0.6 on armel due to an internal compiler error" [High,New]
[17:35] <lool> Oh good thqnks
[17:36] <ogra> https://bugs.launchpad.net/ubuntu/+source/gcc-4.4/+bug/398403/comments/1 is on a lange btw and errors out at the same point the buildd does ...
[17:36] <ubot4> Launchpad bug 398403 in gcc-4.4 "gcc-4.4 fails to build upstart 0.6 on armel due to an internal compiler error" [High,New]
[17:36] <ogra> the confusing part was the testbuild on the other board
[17:36] <ogra> which failed at totally random places
[17:52] <lool> doko: Is there a bug requesting the merge of the patch which broke our binutils bug?
[17:54] <doko> lool: no, binutils doesn't use issues for bug tracking. lets wait for Daniel's reply now