**** BEGIN LOGGING AT Wed Aug 03 02:59:57 2011 Aug 03 08:37:38 morning all Aug 03 08:43:22 what's the compresslevel of the tar.bz2 images created by OE? Aug 03 08:45:38 nevermind, I'll use tar instead. unpacking a tar.bz2 on a beagleboard takes quite some time :) Aug 03 08:51:36 hi bluelightning Aug 03 09:09:30 Does HDX ICAClient work on mini6410? can anyone help me.... does anyone had done that.... Aug 03 09:10:22 I had install on mini6410.... It get connected too.... but getting weird display screen.... Aug 03 09:19:53 sanket: isn't that proprietary? if so this is not the place for support for that, talk to Citrix Aug 03 09:57:41 hi mickey|office Aug 03 09:57:49 good morning pb_ Aug 03 10:43:58 hi all! Aug 03 10:47:50 hi lumag Aug 03 10:48:21 bluelightning, hi, any news regarding meta-hh or meta-opie? Aug 03 10:48:40 lumag: some progress, no release yet Aug 03 10:48:50 in the mean time I fixed some more build issues last night Aug 03 10:49:21 whereabouts is meta-handhelds hosted, by the way? Aug 03 10:49:31 pb_: it's not, that's the issue Aug 03 10:49:36 ah, I see Aug 03 10:49:40 do you want to have it on oe.org? Aug 03 10:50:11 that would be great... previously I was told to put it on github Aug 03 10:50:49 oh, right. who told you that? Aug 03 10:51:55 I don't know of any reason why we couldn't host it at oe.org. if you want to send me the url to a temporary repo that I can pull from, I'll try to set it up later. Aug 03 10:52:08 koen and graeme Aug 03 10:52:27 pb_: ok, great... when I'm in a position to release it I'll give you a shout, thanks Aug 03 10:52:31 ok, cool Aug 03 10:55:49 colleagues, pxa270 boards (armv5te + iwmmxt) seems broken now due to big TUNE patches. Would it be suitable to add iwmmxt tune to oe-core, or it should live in meta-zaurus? Aug 03 10:56:38 I think it would be ok to add to oe-core. Aug 03 10:56:43 Asking about meta-zaurus, as I'm concerned about spitz. Aug 03 10:56:49 pb_, ok Aug 03 10:57:15 obviously I'm not the person who actually decides, but if I were you I would at least try submitting a patch and see what happens Aug 03 11:06:07 ant_work, do you have any recent reports regarding collie? (especially reg. CF cards)? I've tried running several kernels in qemu, but card was not fully detected. Do you have any info? Aug 03 11:13:05 03Josua Mayer  07master * r1a91fc4b70 10openembedded.git/recipes/mplayer/mplayer_git.bb: (log message trimmed) Aug 03 11:13:05 mplayer_git: Add patch to fix linker error on x86 machines Aug 03 11:13:05 For x86 machines the option `--enable-runtime-cpudetection` is passed to the configure script. Aug 03 11:13:05 EXTRA_OECONF_append = " ${@base_contains('MACHINE_FEATURES', 'x86', '--enable-runtime-cpudetection', '',d)} " Aug 03 11:13:05 This causes a linker error like Aug 03 11:13:05 fix undefined references in loader/module.o: Aug 03 11:13:06 module.c:(.text+0xd60): undefined reference to `report_entry' Aug 03 11:14:43 btw: do we have mplayer recipe somewhere for oe-core builds? Aug 03 11:15:44 lumag: hi Aug 03 11:19:11 I sent you te pending patch fixing iwmmxt Aug 03 11:21:36 about cf, here a log http://pastebin.com/se3BMrmU Aug 03 11:21:42 http://pastebin.com/Cfa1VpLe Aug 03 11:21:50 this is in the release branch Aug 03 11:21:56 I think I've seen it before Aug 03 11:23:45 ant_work, strange. I'll have to dig qemu then Aug 03 11:23:59 wait, read this too: http://lists.linuxtogo.org/pipermail/angstrom-distro-users/2011-August/003739.html Aug 03 11:26:22 ant_work, I'm using standard microdrive emulation which works with pxa-based zaurii. StrongARM PCMCIA unit is exactly the same as in PXA. So it's either some generic hw emulation problem (like caches or other stuff) or I don't know. Aug 03 11:26:46 ah, I see Aug 03 11:27:52 ant_work, regarding iwmmxt: this is a bit strange. We already have a tune-iwmmxt.inc, but it IIUC will default for all packages to be built into iwwmxt feed. Aug 03 11:30:03 Most probably we don't want that. I'd instead use armv5te feed/xscale optimization by default and only use iwmmxt for several packages (like mplayer) Aug 03 11:32:17 yes, if we want to share the same feeds Aug 03 11:32:40 are iwmmxt binaries working with recent kernels? Aug 03 11:34:09 lumag: I asked RP but he ws very busy. Maye pb_ will tell us about http://paste.debian.net/124948/ Aug 03 11:37:53 no.... iwmmmxt is still broken with 3.0 Aug 03 11:39:54 ensc|w, any points for bugreport? Aug 03 11:40:19 http://comments.gmane.org/gmane.linux.ports.arm.general/11795 Aug 03 11:41:40 it's probably a bug in the kernel which does not restore iwmmxt registers on task switches Aug 03 11:41:59 (extremly tricky and optimized code) Aug 03 11:43:44 ant_work, pb_ I'd suggest something like this: http://pastebin.de/18080 Aug 03 11:46:36 lumag: seems right if we want to split pxa25x and pxa27x Aug 03 11:46:47 ensc|w, hmmm. Aug 03 11:47:26 ant_work, patch is completely untested Aug 03 11:56:40 lumag: I can reproduced it on a pxa168 in 2 minutes; download https://www.cvg.de/people/ensc/iwmmxt-ld.tar.gz, extract it on tmpfs, chroot into it and execute ./x Aug 03 11:57:01 ensc|w, have you tried pinging Eric Miao directly or via LAKML Aug 03 11:58:51 afair, posting was cc'ed to them Aug 03 12:02:36 ensc|w, also you might be interested in debugging which paths the kernel chooses in kernel/iwmmxt.S, kernel/xscale-cp0.c, etc. Aug 03 12:02:51 lumag: as far as it concerns Zaurus, maybe only mplayer and ffmpeg are concerned Aug 03 12:03:08 stanislav did a bit of work http://www.penguin.cz/~utx/zaurus/software Aug 03 12:03:10 ensc|w, I don't have pxa270 at hand, might have a chance to check on qemu Aug 03 12:03:44 lumag: last year only mplayer http://www.penguin.cz/~utx/zaurus/feed/ipk/iwmmxt/ Aug 03 12:10:40 ant_work, I expected that Aug 03 13:09:05 ant_work, goot collie to boot under qemu Aug 03 13:12:20 Strangely enough the qemu is booting (with lot's of permission denied errors) with default init, but panics right at system prompt with init=/bin/sh Aug 03 13:23:07 ant_work: yeah, that sucks. the issue is that PACKAGE_EXTRA_ARCHS_tune-xx is not (despite what it looks like) an override. Aug 03 13:23:23 see the definition of PACKAGES_EXTRA_ARCHS in bitbake.conf for why setting that variable directly doesn't do what you might expect. Aug 03 13:25:31 btw: what is ??= ? Aug 03 13:27:57 basically like a delayed version of ?=. see the manual. Aug 03 13:43:14 anyone in the OE community going to the desktop summit? Aug 03 14:01:49 not me Aug 03 14:01:54 * pb_ not allowed out very often **** BEGIN LOGGING AT Wed Aug 03 14:05:28 2011 Aug 03 14:06:19 ah, here we go. https://www.desktopsummit.org/program/sessions/degrees-playing-nice-large-companies-open-source Aug 03 14:06:27 I think he is a member of the oe eV at least :-} Aug 03 14:06:35 dirk? Aug 03 14:06:46 I think he mentioend he was going on google+ Aug 03 14:12:51 ant_work, here? Aug 03 14:13:36 yes, back now Aug 03 14:16:10 Regarding collie: I see strange problems when using zImage compiled via OE, but no problems with 2.6.39 compiled with codesourcery gcc. Aug 03 14:23:13 ah Aug 03 14:23:52 I did compile the working 2.6.38 with TC of oe-dev Aug 03 14:24:55 With 2.6.39 compiled with OE I get kernel panic if I set init=/bin/sh right after command prompt Aug 03 14:27:45 iirc we had one bad linaro patch breaking things for arm <7 but that was oe-dev Aug 03 14:30:25 lumag: did yoy test on c7x0 too? Aug 03 14:30:41 ant_work, no. Aug 03 14:31:11 ok, I'll do Aug 03 14:31:34 ant_work, checked tosa via qemu, basic things seem to work. Aug 03 14:31:43 no graphics test ATM. Aug 03 14:31:43 atm oe-core-built kernel boots on poodle/c7x0 Aug 03 14:32:24 ant_work: which linaro patch? someone checked? Aug 03 14:33:00 oh, hrw, that was long ago and maybe khem did it. Should be in the logs Aug 03 14:33:09 ok Aug 03 14:33:26 effect was broken kernel. reported by JaMa iirc Aug 03 15:35:08 luman ?? is weaker than ?= Aug 03 15:35:17 err that was for lumag Aug 03 15:35:39 khem, yeah, got it already. Aug 03 15:35:45 good Aug 03 15:40:07 ant_work, I still get very strange behaviour with collie-qemu & self-compiled kernel. Aug 03 15:40:30 wich gcc? Aug 03 15:40:48 codesourcery. I'll retry with emdebian Aug 03 15:43:56 ant_work, e.g. I can boot console-image with init=/bin/sh, but then I can create files only in /tmp (not in /, e.g.) Aug 03 15:44:07 I get permission denied Aug 03 15:45:51 a few days ago console-image did not get to the prompt in my tests, missing modutils.sh iirc Aug 03 15:46:23 now I've seen koen updated console-image to systemd Aug 03 15:46:31 ant_work, that's with "rw init=/bin/sh" Aug 03 15:47:24 I'll do that soon: I have to check if really devtmpfs let you boot /bin/sh with empty /dev Aug 03 15:47:57 ant_work, yes, it does. Aug 03 15:48:21 ant_work, but then (as it seems to me) you are left w/o ptys: there is no /dev/pts :( Aug 03 15:48:32 ah, I guess console is needed Aug 03 15:50:07 console is working. that's not a problem Aug 03 15:51:11 oh, I had that permission thing fairly recently Aug 03 15:51:26 stupid busybox seemed to be dropping root privs in a situation where it oughtn't have done Aug 03 15:56:06 pb_, seems like it Aug 03 15:56:37 at least I can create files with redirection. Aug 03 15:57:03 via output redirection I mean Aug 03 16:02:28 so seems not kernel issue Aug 03 16:02:54 There seems to be at least one kernel issue, I must recheck it. Aug 03 16:03:16 ok, going now, bbl Aug 03 16:35:02 whats the best way to debug how bitbake is deciding to use a specific package? Aug 03 16:45:34 pb_, do you know any way to disable that spurious root privs drop? Or it's fixed now? Aug 03 16:53:04 03Joshua Lock  07master * rd1f1ebbe50 10bitbake.git/lib/bb/ui/hob.py: Aug 03 16:53:04 bb/ui/hob: save changes to bblayers.conf when using Add Layer menu item Aug 03 16:53:04 Fixes [YOCTO #1283] Aug 03 16:53:04 Signed-off-by: Joshua Lock Aug 03 16:53:14 03Joshua Lock  07master * rb16663e191 10bitbake.git/lib/bb/ui/crumbs/runningbuild.py: (log message trimmed) Aug 03 16:53:14 bb/ui/crumbs/runningbuild: optionally create list entries sequentially Aug 03 16:53:14 In b947e7aa405966262c0614cae02e7978ec637095 Bob changed the behaviour of Aug 03 16:53:14 the RunningBuildModel such that the items are added to the model in a Aug 03 16:53:15 non-sequential order. Aug 03 16:53:15 The messages in the view being listed out of order from how they are Aug 03 16:53:16 received is undesirable for users of the hob UI, therefore this patch adds Aug 03 16:53:16 03Joshua Lock  07master * rd1d6bfab17 10bitbake.git/lib/bb/ui/crumbs/runningbuild.py: (log message trimmed) Aug 03 16:53:17 ui/crumbs/runningbuild: add optional readonly mode, default off Aug 03 16:53:17 In b947e7aa405966262c0614cae02e7978ec637095 Bob started to introduce code Aug 03 16:53:18 for a right-click menu, whilst most of the code is non-invasive it does Aug 03 16:53:18 enable the editable property of the gtk.TreeView which can be confusing. Aug 03 16:53:19 This change adds a readonly parameter, defaulting to False, to the Aug 03 16:53:19 RunningBuildTreeView which if True will prevent the editable property from Aug 03 16:53:59 03Joshua Lock  07master * r972769e636 10bitbake.git/lib/bb/ui/hob.py: Aug 03 16:53:59 bb/ui/hob: show build messages are displayed in the order they're received Aug 03 16:53:59 Use the new sequential option of RunningBuild to ensure this. Aug 03 16:53:59 Fixes the first part of [YOCTO #1311] Aug 03 16:54:00 Signed-off-by: Joshua Lock Aug 03 16:54:00 03Joshua Lock  07master * rd790386112 10bitbake.git/lib/bb/ui/hob.py: Aug 03 16:54:01 bb/ui/hob: disable editing in the build messages tree view Aug 03 16:54:01 Addresses the second part of [YOCTO #1311] Aug 03 16:54:02 Signed-off-by: Joshua Lock Aug 03 16:54:02 03Joshua Lock  07master * r25ec130758 10bitbake.git/lib/bb/ui/crumbs/runningbuild.py: Aug 03 16:54:03 bb/ui/crumbs/runningbuild: emit signal when command fails with exit signal Aug 03 16:54:03 Emit the generic build-complete signal when a command fails with an exit Aug 03 16:54:59 signal enabling the UI to update itself accordingly. Aug 03 16:54:59 Addresses [YOCTO #1265] Aug 03 16:54:59 Signed-off-by: Joshua Lock Aug 03 17:27:24 lumag: ping Aug 03 17:27:33 lumag: about CF on Collie Aug 03 17:27:59 lumag: there was letter in angstrom-devel ML that have similar issues Aug 03 17:28:31 Jay7, I got my collie-qemu to boot from CF. But now I got another problem. Now with busybox, which looks like it drops root privileges incorrectly. Aug 03 17:29:07 Strangely enough it affected only collie build. Aug 03 17:30:02 lumag: http://lists.linuxtogo.org/pipermail/angstrom-distro-users/2011-July/003737.html Aug 03 17:30:14 and http://lists.linuxtogo.org/pipermail/angstrom-distro-users/2011-August/003739.html Aug 03 17:30:40 lumag: is login does not working? Aug 03 17:31:00 no. It's stuck before that Aug 03 17:31:09 iirc, this was most known issue with bb :) Aug 03 17:31:38 * lumag is not a bb fan. Aug 03 17:33:29 Currently I'm rebuilding collie image. If that won't help, I'll try rebuilding busybox. Aug 03 17:33:38 But currently the image is unusable :( Aug 03 17:34:34 Got to go now. Aug 03 17:34:36 Bye! Aug 03 17:34:39 bb Aug 03 19:10:31 msm: bitbake -g will emit the dep graph Aug 03 19:10:46 msm: then you can look around in the .dot files Aug 03 19:11:31 khem: im only getting one proivider for virtual/kernel for my machine Aug 03 19:11:39 you can also use the depexp ui Aug 03 19:11:43 i can't seem to find why it's not including linux-yocto Aug 03 19:11:52 yeah depexp ui Aug 03 19:12:06 msm: which machine is it Aug 03 19:12:11 my own Aug 03 19:12:16 so i need to find out what im missing Aug 03 19:12:43 msm: hmmm linux-yocto recipes should know about your machine first Aug 03 19:12:44 i've made a linux-yocto_3.0.bbappend Aug 03 19:13:11 and my preferred providor is linux-yocto Aug 03 19:13:36 what does it show as provider Aug 03 19:14:03 just linux dummy Aug 03 19:14:05 if its not in COMPATIBLE_MACHINE for linux-yocto then it wont show up Aug 03 19:14:20 let me look Aug 03 19:14:25 or something equivalent of COMPATIBLE_MACHINE Aug 03 19:14:27 that was in the bbappend file at one point Aug 03 19:14:47 i had the line commented out… but i had it at one point Aug 03 19:14:49 msm: its a regexp Aug 03 19:15:04 so you might have to append to it Aug 03 19:15:40 COMPATIBLE_MACHINE += "(p2020ds)" Aug 03 19:15:48 has that format changed any? Aug 03 19:16:20 no Aug 03 19:16:25 COMPATIBLE_MACHINE = "(qemuarm|qemux86|qemuppc|qemumips|qemux86-64)" Aug 03 19:16:38 is what I see in original .bb Aug 03 19:16:55 right, some older linux-yocto.bbappend has the += format i pasted Aug 03 19:17:30 depexp… nice =) Aug 03 19:18:22 just COMPATIBLE_MACHINE_ = "yourmachine Aug 03 19:18:24 " Aug 03 19:18:35 in your .append file Aug 03 19:18:38 i see that notation as well Aug 03 19:18:39 ok Aug 03 19:19:37 you might also need KMACHINE for your machine Aug 03 19:19:39 still just picked up Aug 03 19:19:43 linux-dummy Aug 03 19:19:55 i have KMACHINE_p2020ds= "yocto/standard/fsl-p2020ds" Aug 03 19:19:58 ok Aug 03 19:20:04 not sure what that yocto/standard rferrs too though Aug 03 19:20:10 what does linux-dummy look like Aug 03 19:20:38 its just the most basic recipe Aug 03 19:20:51 provides virtual/kernel - but does nothing Aug 03 19:22:51 PREFERRED_PROVIDER_virtual/kernel ?= "linux-yocto" Aug 03 19:23:10 is that somewhere in your conf metadata Aug 03 19:23:12 i have an '=' Aug 03 19:23:18 yes Aug 03 19:23:23 PREFERRED_PROVIDER_virtual/kernel = "linux-yocto" Aug 03 19:23:23 PREFERRED_VERSION_virtual/kernel = "3.0" Aug 03 19:23:50 where do u have it ? Aug 03 19:23:53 in local.conf Aug 03 19:24:01 in machine.conf Aug 03 19:24:10 ok Aug 03 19:24:27 what is this path refer too: yocto/standard/fsl-mpc8315e-rdb Aug 03 19:24:34 in KMACHINE Aug 03 19:24:43 I dont have much idea Aug 03 19:24:57 but its kind of machine specific metadata Aug 03 19:25:01 for that kernel Aug 03 19:25:12 may be .config and patches if any I believe Aug 03 19:25:35 where does this live in the tree? Aug 03 19:27:23 i just can't see where else to provide an override for my new machine Aug 03 19:27:33 grepping through the source for mpc8315 for comparisoon Aug 03 19:28:06 its in kernel sources I think Aug 03 19:28:20 but it should first select the right kernel Aug 03 19:28:28 can you do this Aug 03 19:28:39 bitbake -e virtual/kernel > /tmp/xx Aug 03 19:28:44 and paste xx somewhere Aug 03 19:29:03 yep Aug 03 19:34:54 http://pastebin.com/8Vubys1t Aug 03 19:34:58 still looking at this myslf Aug 03 19:45:49 msm: COMPATIBLE_MACHINE=None Aug 03 19:45:53 this is the problem I think Aug 03 19:46:08 hmm Aug 03 19:46:27 i added my machine to COMPATIBLE_MACHINE in linux-yocto too and it worked Aug 03 19:47:17 msm: ok that means your .bbappend was not really being parsed in Aug 03 19:47:44 i have an idea Aug 03 19:47:47 how does your bbappend look like Aug 03 19:48:04 which makes me look silly if it works Aug 03 19:48:08 do bitbake -DDDD virtual/kernel and paste the output Aug 03 19:48:53 i have to add bbppaned files to BBFILES? Aug 03 19:49:12 bbappend* Aug 03 19:49:27 well yes Aug 03 19:49:33 atleast the search path Aug 03 19:49:56 and… it seems to be working Aug 03 19:50:15 BBFILES += ${LAYERDIR}/recipes-*/*.bbappend" Aug 03 19:50:16 khem: thanks for the assistance Aug 03 19:50:20 np Aug 03 19:50:47 is virtual/kernel selected by most images as well? Aug 03 19:50:57 i dont need to add that as a package somewhere to build it for this machine? Aug 03 19:51:11 msm: usually virtual/kernel is decided by machine Aug 03 19:52:03 yes, by machine config or some inc required by machine config Aug 03 19:55:48 fray: around? Aug 03 20:52:04 hi Jay7 Aug 03 20:52:14 gnutoo: hi Aug 03 20:52:38 Jay7: you were involved in kexecboot right? Aug 03 20:52:46 gnutoo: yes, sure Aug 03 20:53:02 my question is: can you pass atags to a kexeced kenrel Aug 03 20:53:09 for instance from /proc/atags Aug 03 20:53:46 hm.. shouldn't atags be in /proc/cmdline in the end? Aug 03 20:54:01 no it's a different thing Aug 03 20:54:25 can you show contents of /proc/atags? Aug 03 20:54:33 ok Aug 03 20:55:29 I guess it contains not only the cmdline Aug 03 20:55:33 but also other stuff Aug 03 20:55:35 right? Aug 03 20:56:01 seems I don't know.. Aug 03 20:56:28 afaik, it was just other way to pass options to kernel Aug 03 20:56:31 from bug20: Aug 03 20:56:34 http://www.pastie.org/2316308 Aug 03 20:57:17 last part looks like just kernel cmdline options Aug 03 20:57:31 not sure about first part (AT....) Aug 03 20:57:52 let me find other atags Aug 03 20:57:56 better to ask some kernel guys :) Aug 03 20:58:01 (other device) Aug 03 21:01:06 I start to remember Aug 03 21:01:19 I think there is also the board arch number Aug 03 21:01:40 all that I know is that it's the interface between the bootloader and the linux kenrel Aug 03 21:02:13 web is very slow here Aug 03 21:02:21 (far wifi AP + VPN) Aug 03 21:03:10 uboot documentation also say: Aug 03 21:03:44 The maximum clock rate allowed depends on the silicon populated on the EVM.[...]This information is Aug 03 21:03:47 passed to the Linux kernel using the ATAG_REVISION atag. Aug 03 21:04:28 arm/Booting seem to have something too in the linux kenrel documentation Aug 03 21:05:56 so it seem a lot more Aug 03 21:06:24 so who knows how atag works with kexecboot? Aug 03 21:06:31 or rather who could know htat Aug 03 21:06:33 *that Aug 03 21:07:00 gnutoo: s/kexecboot/kernel/ :) Aug 03 21:07:36 kexecboot does not pass atags to next kernel Aug 03 21:07:51 for the kenrel I know Aug 03 21:07:52 but you may create boot.cfg with that options in APPEND Aug 03 21:08:11 as quick workaround Aug 03 21:08:21 I know more or less(less than more) how it works at first boot Aug 03 21:08:40 but I've no idea why there is no atag options in kexec Aug 03 21:08:47 (the kexec program) Aug 03 21:09:11 gnutoo: if atags just other way to specify kernel cmdline, it have no sense Aug 03 21:09:23 just add options to kexec --command-line or --append Aug 03 21:09:26 as I just said it's more than that Aug 03 21:09:35 it also specify machine number Aug 03 21:09:38 and a lot of stuff Aug 03 21:09:55 my real problem is that since I made kexecboot work with n900 Aug 03 21:09:58 well.. then it's out of my knowledge.. Aug 03 21:10:07 an user complained that many stuff didn't work anymore in maemo Aug 03 21:10:24 + the fact that there is one minute delay between kexec -e and the kernel starting Aug 03 21:10:31 I don't know how to pass atags to kexec'ed kernel.. Aug 03 21:10:40 so I wanted to try passing atags Aug 03 21:10:43 to see..... Aug 03 21:10:49 ok Aug 03 21:10:58 ant should know? right? Aug 03 21:11:04 but he's not here right now Aug 03 21:11:35 gnutoo: he may know more but I'm unsure about details :) Aug 03 21:11:41 ok Aug 03 21:11:53 may be larsc on #qi-hw may say something Aug 03 21:12:05 but he is mips hacker, not arm Aug 03 21:12:13 and iirc atags are arm-specific Aug 03 21:12:14 he's an arm hacker too Aug 03 21:12:25 he does the freerunner kernel too Aug 03 21:12:29 try to bother him then :) Aug 03 21:12:35 there is a preserve context option Aug 03 21:12:38 what does it do? Aug 03 21:13:04 what do you mean? Aug 03 21:13:15 kexec has: Aug 03 21:13:34 --load-preserve-context Aug 03 21:14:05 the help says: Load the new kernel and preserve context of current kernel during kexec Aug 03 21:15:54 ah Aug 03 21:15:58 seems I understand Aug 03 21:16:12 it does not free'd previous kernel memory e.g. Aug 03 21:16:21 so you may reuse/dump it Aug 03 21:16:30 https://lists.linux-foundation.org/pipermail/linux-pm/2008-March/016953.html Aug 03 21:20:53 ok Aug 03 21:20:56 thanks Aug 03 22:34:43 khem: that yocto/standard/xyz var is actually a branch on the linux tree i believe Aug 03 22:35:24 KMACHINE_p2020ds = "linux-yocto-branch-goes-here" Aug 04 01:47:09 Did linuxtogo just change ML stuff? Aug 04 01:47:21 Or is tbird being upset for some reason, all oe-core mails aren't being filtered for me Aug 04 01:47:36 Er, nm, -ENOTPAYINGATTENTION **** ENDING LOGGING AT Thu Aug 04 02:59:57 2011