**** BEGIN LOGGING AT Mon Apr 23 02:59:58 2012
Apr 23 07:21:31 <mckoan> good morning
Apr 23 07:34:15 <likewise> gm all
Apr 23 07:48:57 <_tasslehoff_> I've got a new work pc. Quadcore (i7-3820 @ 3.60GHz), 8GB RAM and SSD. Now I'm pondering about sensible starting points for BB_NUMBER_THREADS and PARALLEL_MAKE for building oe-core. Any advice?
Apr 23 07:50:04 <_tasslehoff_> Asked this before, but then I was on a crappy laptop with 4GB RAM and an HDD, so I figure things may have changed.
Apr 23 07:51:29 <mckoan> _tasslehoff_: BB_NUMBER_THREADS=8 PARALLEL_MAKE>=4
Apr 23 07:52:26 <mckoan> _tasslehoff_: anyway you could call : time bitbake xxx , to test how long it takes
Apr 23 07:52:51 <_tasslehoff_> mckoan: thanks. that's the plan.
Apr 23 08:02:23 <Zagor> I'd run BB=8 PM=8.
Apr 23 08:02:48 <Zagor> the i7-3820 has 8 threads, after all
Apr 23 08:26:31 <ant_work> JaMa|Wrk: ping
Apr 23 08:26:40 <JaMa|Wrk> pong
Apr 23 08:26:53 <ant_work> hi
Apr 23 08:27:19 <ant_work> do you remember klibc vs kernel headers > 3.2 issue?
Apr 23 08:27:25 <JaMa|Wrk> yes
Apr 23 08:27:50 <ant_work> I 'fixed' the build with 4 symlinks but that's an hack
Apr 23 08:28:19 <ant_work> do you think we need to tweak the headers installation?
Apr 23 08:28:52 <JaMa|Wrk> I had similar hack to fix it, but IIRC you said that with newer klibc it works OOTB
Apr 23 08:29:18 <ant_work> eh, I thought...this happens with klibc master too, not only 1.5.25
Apr 23 08:29:29 <ant_work> they fixed another error
Apr 23 08:29:52 <JaMa|Wrk> hmm
Apr 23 08:30:28 <ant_work> it look slike make headers_install doeas not install /generated/asm
Apr 23 08:30:30 <JaMa|Wrk> and btw 3.4-rc* are not booting.. testing rc4 now, but few commits back in rc3 it was still failing
Apr 23 08:30:54 <ant_work> oh, for that I'm waiting yocto-3.4 ;)
Apr 23 08:33:58 <JaMa|Wrk> good luck with that
Apr 23 08:35:27 <ant_work> Apr 08 21:38:29 <dromede_> i also managed to boot 3.4-rc2
Apr 23 08:35:50 <ant_work> Apr 08 21:40:16 <dromede_> 3.4-rc2 craps out at sharpsl_save_param() in arch/arm/common/sharpsl_param.c
Apr 23 08:36:02 <ant_work> Apr 08 21:40:28 <dromede_> i have no idea why
Apr 23 08:36:08 <ant_work> Apr 08 21:40:50 <dromede_> i commented the function out and then it boots
Apr 23 08:36:47 <ant_work> that was my last status-update
Apr 23 08:45:45 <jonasdn> hi bluelightning 
Apr 23 08:46:11 <bluelightning> hi jonasdn, all
Apr 23 08:46:19 <jonasdn> want to take a look at an attempt at adding sysvinit for udev-182, a naive first approach, but works for me atm
Apr 23 08:46:32 <jonasdn> http://pastebin.com/3Ehb4caf
Apr 23 08:46:54 * bluelightning looks
Apr 23 08:50:34 <likewise> hi bluelightning, all
Apr 23 08:51:34 <likewise> _tasslehoff_: SSD alone gave me 30% boost, and that was with an Intel SSD from two years ago
Apr 23 08:53:00 <bluelightning> jonasdn: I'm no expert on systemd but it looks like there are a number of changes in here that aren't directly related to enabling systvinit...
Apr 23 08:56:38 <jonasdn> I'm probably less of an expert on systemd. What changes are you thinking of? The reason that the init script was removed in later version of udev recipes was that the location of the binaries changed and the script didn't work anymore, so that is what I tried to solve.
Apr 23 08:56:58 <jonasdn> but It's icky
Apr 23 08:57:14 <jonasdn> and I guess something more structured is needed in oe to deal with sysvinit/systemd-partitioning
Apr 23 09:02:41 <ant_work> jonasdn: at the moment the situation is clear
Apr 23 09:02:49 <ant_work> oe-core -> sysvinit
Apr 23 09:02:57 <ant_work> meta-oe -> systemd
Apr 23 09:03:05 <jonasdn> ok
Apr 23 09:03:20 <jonasdn> that's sort of what I wanted to know, that is a clear statement?
Apr 23 09:03:24 <ant_work> most distros using meta-oe have moved to systemd (Angstrom shr iirc)
Apr 23 09:03:48 <ant_work> meta-oe is still seen as a pandora-box 
Apr 23 09:03:55 <ant_work> situation will improve hopefully soon
Apr 23 09:04:04 <jonasdn> yeah :)
Apr 23 09:04:20 <jonasdn> we ran into trouble with udev-182, we have a DISTRO_FEATURES like:
Apr 23 09:04:26 <jonasdn> DISTRO_FEATURES = "argp ext2 largefile ${DISTRO_FEATURES_LIBC}"  
Apr 23 09:04:31 <jonasdn> and we don't include systemd
Apr 23 09:04:39 <jonasdn> but udev-182 is still used when using meta-oe
Apr 23 09:04:42 <jonasdn> so udev just didn't start :)
Apr 23 09:04:55 <jonasdn> since it doesn't have sysvinit start scripts
Apr 23 09:04:55 <ant_work> just remove it from the layer
Apr 23 09:05:03 <ant_work> it = udev
Apr 23 09:05:17 <jonasdn> yeah, that is what we do, we will use PREFERRED_VERSION_udev = "173" I think
Apr 23 09:05:24 <jonasdn> or 164 since 173 might be removed from meta-oe 
Apr 23 09:05:28 <ant_work> you'll get the 1.64 of oe-core
Apr 23 09:05:47 <jonasdn> just wanted to know if it was a clear statement that for meta-oe systemd is required
Apr 23 09:06:06 <ant_work> heh
Apr 23 09:06:17 <jonasdn> it's bothersome tho, because 164 is well, not new
Apr 23 09:06:26 <jonasdn> if we want bugfixes or something we are forced to systemd
Apr 23 09:06:38 <jonasdn> I guess we will have to embrace it :)
Apr 23 09:06:49 <ant_work> I think that next week after 1.2 release many patches will come
Apr 23 09:07:05 <ant_work> udev upgrade was already discussed on the ML iirc
Apr 23 09:07:26 <jonasdn> so maybe a oe-core udev-182 version will come, with sysvinit support?
Apr 23 09:07:34 <jonasdn> or something higher than 164 at least
Apr 23 09:07:39 <ant_work> sure
Apr 23 09:07:46 <jonasdn> nifty
Apr 23 09:07:54 <ant_work> fwiw there are other issues in meta-oe, at least in my experience
Apr 23 09:08:10 <ant_work> notably with xserver-nodm-init
Apr 23 09:15:09 <jonasdn> thank you
Apr 23 09:18:34 <bluelightning> that's not an official statement
Apr 23 09:18:57 <bluelightning> we do want meta-oe to not make any unexpected changes
Apr 23 09:34:58 <ant_work> yeah, there isn't any official statement, the interaction between oe-core and meta-oe is slowly evoluting atm
Apr 23 10:04:34 <ant_work> pb_: ping
Apr 23 10:08:16 <_tasslehoff_> do you think there
Apr 23 10:08:21 <_tasslehoff_> oops
Apr 23 10:09:01 <_tasslehoff_> do you think there's a big difference between using (Ubuntu) 64-bit and 32-bit w/PAE-kernel when compiling oe-core?
Apr 23 10:09:44 <XorA> nope
Apr 23 10:12:14 <_tasslehoff_> then I'm tempted to install 32-bit after all, since it seems our codebase isn't exactly 64-bit ready...
Apr 23 10:13:11 <Zagor> will you be able to use all 8GB of ram with a 32-bit kernel?
Apr 23 10:14:27 <Zagor> oh, PAE. I need lunch... :)
Apr 23 10:40:07 <ericbutters> i got angstom with lxde running on my arm device. i can login on serial console. but if i call some x application and did not export DISPLAY=:0 the console hangs (not receiving any characters anymore) http://pastebin.com/raw.php?i=xc9rG3j7
Apr 23 10:40:20 <ericbutters> anyone can say what can cause that?
Apr 23 10:44:21 <ericbutters> ok. it is always that i can not enter commands after i logged in. the time where it happens is different.. 
Apr 23 11:08:08 <jonasdn> so BB_GENERATE_TARBALLS = "1" makes sure I get tarballs for the dowloaded packages, is there a way to set where they will be generated to, or is it allways the download/-directory?
Apr 23 12:31:58 <ericbutters> is there any machine support for S5PC100 ?
Apr 23 12:42:42 <GNUtoo-desktop> ericbutters, hi no idea about s5pc100 but there is s5pc110
Apr 23 12:43:03 <GNUtoo-desktop> in meta-smartphone/meta-samsung
Apr 23 12:50:41 <ericbutters> GNUtoo-desktop ah nice! thanks.. another question: i have my custom kernel as a git repostory on my local harddrive. how to modify SRC_URI from SRC_URI = "git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung.git;protocol=git;tag=v3.2-rc5" to access that local repository
Apr 23 12:51:02 <ericbutters> ?
Apr 23 12:51:35 <GNUtoo-desktop> I don't know, maybe host the repo somewhere
Apr 23 12:52:02 <GNUtoo-desktop> back in the oe-classic days it could be done with 2 classes
Apr 23 12:53:21 <GNUtoo-desktop> srctree and  gitver
Apr 23 12:53:40 <ericbutters> what if the board_defconfig i setup did not enable USB? are the tools included into the rootfs image build without USB?
Apr 23 12:54:20 <GNUtoo-desktop> no, only your kernel will lack USB, but....
Apr 23 12:54:38 <ericbutters> is the defconfig of the linux kernel or the linux-kernel itself responible for the rootfs tools?
Apr 23 12:54:44 <GNUtoo-desktop> if you don't put usb in machine features in your machine config....they will not go in your image
Apr 23 12:54:53 <GNUtoo-desktop> no it's not
Apr 23 12:54:56 <ericbutters> ok
Apr 23 12:55:28 <GNUtoo-desktop> machine config means the thing that is in conf/machine/
Apr 23 12:57:09 <bluelightning> ericbutters: try externalsrc.bbclass - that's what that is designed for
Apr 23 12:57:22 <bluelightning> (using a local source checkout)
Apr 23 12:57:25 <ericbutters> my problem is: i have lxde (angstrom) image .. that boots.. but usb is not working. tested my linux-kernel with other rootfs (buildroot) here usb is working and i can use mouse. and after a while the serial console is not working anymore (angstrom lxde) ?
Apr 23 12:57:56 <ericbutters> do you know such problems?
Apr 23 12:58:54 <ericbutters> bluelightning where can i find syntax example for externalsrc.bbclass?
Apr 23 12:59:30 <bluelightning> ericbutters: there are some brief instructions at the top of that class file
Apr 23 13:02:22 <ericbutters> could you pls take a quick look at my image configuration here: http://pastebin.com/raw.php?i=rmXaqyui
Apr 23 13:02:33 <ericbutters> do you something i miss? 
Apr 23 13:02:39 <ericbutters> some problems?
Apr 23 13:04:52 <ericbutters> i did: MACHINE_FEATURES = "screen" in my conf/machine/smdk2440.conf .. but maybe that is the problem
Apr 23 13:04:57 <ericbutters> should be MACHINE_FEATURES = "kernel26 screen" ?
Apr 23 13:05:37 <bluelightning> ericbutters: no, kernel26 is obsolete
Apr 23 13:05:40 <GNUtoo-desktop> it's an s3c2440?
Apr 23 13:07:53 <ericbutters> no in fact it is s5pc100
Apr 23 13:09:05 <ericbutters> bluelightning what replaces kernel26? i use v3.2
Apr 23 13:09:30 <bluelightning> ericbutters: nothing; it was only necessary in the days when it was necessary to distinguish between 2.4 and 2.6 kernels
Apr 23 13:09:51 <bluelightning> now we dropped support for 2.4 it doesn't do anything
Apr 23 13:10:07 <ericbutters> so it is ok to say MACHINE_FEATURES = "screen" ?
Apr 23 13:10:18 <bluelightning> assuming you want to enable no other features, yes
Apr 23 13:10:35 <ericbutters> but this has no affect on lxde?
Apr 23 13:10:40 <ericbutters> effect?
Apr 23 13:11:08 <bluelightning> I wouldn't think so, but "git grep MACHINE_FEATURES" will tell you definitively
Apr 23 13:12:21 <ericbutters> so OE build linux-kernel is totaly independent  of rootfs-image built?
Apr 23 13:12:54 <GNUtoo-desktop> ericbutters, if lxde uses tasks it could have effects
Apr 23 13:15:00 <ericbutters> ok.. i think i neet to add "usbhost" to MACHINE_FEATURES
Apr 23 13:25:49 <ericbutters> ok one thing: i changed my file under conf/machine .. if i call bitbake again, these changes are uses?
Apr 23 13:25:55 <ericbutters> used? sorry ;)
Apr 23 13:26:16 <ericbutters> or do i need to rm -rf tmp/ ?
Apr 23 13:27:02 <otavio> Hi :)
Apr 23 13:27:27 <otavio> Can people test u-boot updating and report it on mailing list?
Apr 23 13:31:56 <bluelightning> ericbutters: it depends; if you have OEBasicHash as your BB_SIGNATURE_HANDLER then a change to any variable will result in a rebuild of anything that depends on that variable
Apr 23 13:32:26 <bluelightning> otherwise, for variables in your machine configuration you likely need to figure out what you need to rebuild yourself or just rebuild everything
Apr 23 13:33:10 <bluelightning> like I mentioned above though git grep MACHINE_FEATURES will show where behaviour is being influenced by that variable
Apr 23 16:04:56 <tasslehoff> Anyone know where the OMAPES variable that libgles-omap3 uses is set? I've been grepping TMPDIR and the recipes, but can't figure it out. Beagleboard gets a -L<something>/libgles-omap3-4.05.00.03-r25/gfx_rel_es3.x while my board gets -L<something>/libgles-omap3-4.05.00.03-r25/gfx_rel, and it seems to be the OMAPES envvar that controls it.
Apr 23 16:16:55 <tasslehoff> hm. I guess the Q is how the ${BINLOCATION} in TARGET_CC_ARCH in libgles-omap3.inc, has gotten an _es<n>.x suffix when the compilation starts.
Apr 23 17:18:49 <ericbutters> depmod: QM_MODULES: Function not implemented
Apr 23 17:19:02 <ericbutters> alignment: ignoring faults is unsafe on this CPU.  Defaulting to fixup mode.
Apr 23 17:20:49 <ericbutters> the serial console still hangs up after login.. anyone?
Apr 23 19:25:28 <fray> how do I stop a particular package from building using parallel make?
Apr 23 19:26:19 <tasslehoff> what will the ?= do? set something if it is not already set?
Apr 23 19:26:41 <fray> foo ?= bar  -- set foo to bar if foo has not been previous set..
Apr 23 19:26:55 <fray> foo ??= bar  -- wait until the very end of evaluation, and if foo is STILL not set, then set it to bar
Apr 23 19:29:26 <tasslehoff> fray: thanks. where should I have read this instead of asking you? :)
Apr 23 19:29:34 <tasslehoff> The Source?
Apr 23 19:31:24 <fray> the bitbake manual (look in the bitbake directory) has that..
Apr 23 19:31:46 <fray> it lists the various evaluation items.. +=, .=, =+, =., =, :=, ?=, and ??=
Apr 23 19:31:55 <fray> (and of course "=")
Apr 23 19:33:03 <tasslehoff> ah, didn't know the docs was in there. I've just seen outdated versions online :)
Apr 23 19:44:38 <tasslehoff> to answer my own Q from earlier. my layer has an omap3.inc, but I missed "require conf/machine/include/soc-family.inc", and got BINLOCATION instead of BINLOCATION_omap3. case closed.
Apr 23 21:21:46 <GNUtoo-desktop> bluelightning, hi, by what is palmtopdir set?
Apr 23 21:22:04 <bluelightning> GNUtoo-desktop: it's now set via the palmtop-defs bbclass
Apr 23 21:22:44 <bluelightning> GNUtoo-desktop: it used to be in bitbake.conf but Opie-specific stuff has no place in OE-core's bitbake.conf really, so I moved them out
Apr 23 21:23:11 <GNUtoo-desktop> ok
Apr 23 21:23:12 <bluelightning> so every recipe that refers to palmtopdir/palmqtdir now inherits palmtop-defs
Apr 23 21:23:19 <GNUtoo-desktop> because for me it's not set in that recipe:
Apr 23 21:23:20 <bluelightning> (every recipe in meta-opie at least)
Apr 23 21:23:41 <GNUtoo-desktop> opie-init.inc
Apr 23 21:24:13 <GNUtoo-desktop> which leads to opie not starting for me because it looks in the wrong dir
Apr 23 21:24:50 <bluelightning> GNUtoo-desktop: oh dear... you're right :/
Apr 23 21:24:55 <bluelightning> how did that get missed?
Apr 23 21:24:58 <bluelightning> I'll fix it now
Apr 23 21:25:22 <GNUtoo-desktop> ok
Apr 23 21:25:35 <GNUtoo-desktop> thanks a lot for fixing
Apr 23 21:25:47 <bluelightning> oh I know why
Apr 23 21:26:17 <bluelightning> my original searches for recipes referring to palmtopdir excluded files containing "inherit opie"
Apr 23 21:26:24 <bluelightning> which that file does, except it's commented out :(
Apr 23 21:26:29 <GNUtoo-desktop> ok
Apr 23 21:26:58 <GNUtoo-desktop> btw not related, but is it possible to get a bigger quit button?
Apr 23 21:27:04 <GNUtoo-desktop> like by changing some settings
Apr 23 21:28:57 <msm> can anyone comment if builds work over NFS?
Apr 23 21:31:47 <bluelightning> GNUtoo-desktop: ok, fix pushed
Apr 23 21:31:57 <bluelightning> GNUtoo-desktop: you mean the close button in the top right?
Apr 23 21:32:01 <GNUtoo-desktop> yes
Apr 23 21:32:24 <bluelightning> that's controlled by the theme... it's somewhat annoying that it doesn't adapt to the screen size for any of the themes though
Apr 23 21:32:31 <bluelightning> it's something I'd like to fix
Apr 23 21:33:17 <GNUtoo-desktop> ok
Apr 23 21:33:41 <GNUtoo-desktop> because the right part is not very sensitive for my screen
Apr 23 21:33:45 <GNUtoo-desktop> I'll try to recalibrate
Apr 23 21:34:19 <bluelightning> I think some themes have larger buttons than others
Apr 23 21:34:36 <GNUtoo-desktop> ok
Apr 23 21:42:01 <bluelightning> msm: it might work, but I doubt we would consider it officially supported
Apr 23 21:42:18 <bluelightning> if you're talking about TMPDIR being on NFS
Apr 23 21:46:29 <GNUtoo-desktop> hmmm ts_harvest don't recognize one of the final touch....
Apr 23 21:47:02 <msm> bluelightning: i seem to have issues
Apr 23 21:47:11 <msm> bluelightning: its just stuck at 0% on an older release
Apr 23 21:47:18 <msm> anyways, just taking notes
**** ENDING LOGGING AT Tue Apr 24 02:59:58 2012