Re: [OE-core] Where is atom-pc.conf hiding?
Op 13 jun 2011, om 23:30 heeft Richard Purdie het volgende geschreven: On Mon, 2011-06-13 at 22:36 +0200, Koen Kooi wrote: Op 13 jun 2011, om 22:28 heeft Saul Wold het volgende geschreven: On 06/13/2011 11:31 AM, Koen Kooi wrote: Hi, Khem was asking if I could reproduce the recent x86 breakage he was seeing[1] and I ran into another bug: koen@dominion:/OE/tentacle/sources/meta-intel$ git blame meta-n450/conf/machine/n450.conf | grep atom 158f88d7 (Saul Wold 2011-01-03 15:33:52 -0800 6) require conf/machine/atom-pc.conf meta-yocto seems to be the place you need to look! I hope that the layering tools can help to detect and inform folks of this like of dependency. Isn't meta-yocto supposed to a the integration layer with no new parts? I can't use meta-yocto since it has conflicting beagleboard stuff in it, which means that meta-intel is now broken for me as well. That surely isn't the intended plan?!?! The plan on public record is that atom-pc moves to meta-intel as soon as the layer tooling comes online and meta-yocto becomes its own repo (which at present its not but its certainly the intent). So can someone at least put that in the meta-intel or meta-n450 README? Most layers now have a README with the dependency info (e.g. http://git.shr-project.org/git/?p=meta-smartphone.git;a=blob;f=meta-palm/README;h=200a3c83ff11ad790fe4e03e93a26520ec0c6714;hb=HEAD ) It would be nice if the ones on yocto-project.org had somethign similar :) ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On Tue, 2011-06-14 at 08:44 +0100, Phil Blundell wrote: On Mon, 2011-06-13 at 15:10 -0700, Tom Rini wrote: Until then, and even afterwards can we please get some testing of non-poky builds done? I know the autobuilder is full but can't we toss a few things onto a personal box and try that a few times a week? I can probably find some spare cpu cycles to do testing. Is there an existing autobuild/autotest infrastructure that we can conveniently use to drive the tests and report the status? Buildbot is what Yocto is using and the documentation/sample config for what we do is available at: http://git.yoctoproject.org/cgit.cgi/poky-autobuilder/ Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On Tue, 2011-06-14 at 00:40 -0700, Koen Kooi wrote: Op 13 jun 2011, om 23:30 heeft Richard Purdie het volgende geschreven: On Mon, 2011-06-13 at 22:36 +0200, Koen Kooi wrote: Op 13 jun 2011, om 22:28 heeft Saul Wold het volgende geschreven: On 06/13/2011 11:31 AM, Koen Kooi wrote: Hi, Khem was asking if I could reproduce the recent x86 breakage he was seeing[1] and I ran into another bug: koen@dominion:/OE/tentacle/sources/meta-intel$ git blame meta-n450/conf/machine/n450.conf | grep atom 158f88d7 (Saul Wold 2011-01-03 15:33:52 -0800 6) require conf/machine/atom-pc.conf meta-yocto seems to be the place you need to look! I hope that the layering tools can help to detect and inform folks of this like of dependency. Isn't meta-yocto supposed to a the integration layer with no new parts? I can't use meta-yocto since it has conflicting beagleboard stuff in it, which means that meta-intel is now broken for me as well. That surely isn't the intended plan?!?! The plan on public record is that atom-pc moves to meta-intel as soon as the layer tooling comes online and meta-yocto becomes its own repo (which at present its not but its certainly the intent). So can someone at least put that in the meta-intel or meta-n450 README? Most layers now have a README with the dependency info (e.g. http://git.shr-project.org/git/?p=meta-smartphone.git;a=blob;f=meta-palm/README;h=200a3c83ff11ad790fe4e03e93a26520ec0c6714;hb=HEAD ) It would be nice if the ones on yocto-project.org had somethign similar :) _ Would something like this do the trick for now (if so, I can do similar for the other meta-intel bsps)? From ae612a895b320471f65e289328fd351d4094a7c2 Mon Sep 17 00:00:00 2001 Message-Id: ae612a895b320471f65e289328fd351d4094a7c2.1308077431.git.tom.zanu...@intel.com From: Tom Zanussi tom.zanu...@intel.com Date: Tue, 14 Jun 2011 13:49:24 -0500 Subject: [PATCH] meta-n450: add dependency info Add information on the dependencies needed to use the meta-n450 layer. Signed-off-by: Tom Zanussi tom.zanu...@intel.com --- meta-n450/ReleaseNotes | 30 ++ 1 files changed, 30 insertions(+), 0 deletions(-) diff --git a/meta-n450/ReleaseNotes b/meta-n450/ReleaseNotes index 23ac815..4257d94 100644 --- a/meta-n450/ReleaseNotes +++ b/meta-n450/ReleaseNotes @@ -1,8 +1,38 @@ Poky Laverne Release 4.0 +- Dependencies - Features and Updates - Known Issues +DEPENDENCIES + + +This layer depends on: + + URI: git://git.openembedded.org/bitbake + branch: master + + URI: git://git.openembedded.org/openembedded-core + layers: meta + branch: master + + URI: git://git.yoctoproject.org/poky + layers: meta-yocto + branch: master + +NOTE: n450 currently depends on meta-yocto due to atom-pc +dependencies. The plan is to move atom-pc to meta-intel as soon as +the layer tooling comes online and meta-yocto becomes its own repo. +At that point, meta-yocto will no longer be required. + +Note also that the layering with respect to the poky repo is in +transition; it currently contains both meta from openembedded-core and +bitbake along with some other files and directories which will +eventually be contained in standalone repos once the ongoing layering +work has been completed. As such, the master branch of the poky repo +can at the present time be used to satisfy all meta-intel +dependencies. + FEATURES AND UPDATES -- 1.7.0.4 __ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
Op 14 jun 2011, om 22:20 heeft Tom Zanussi het volgende geschreven: On Tue, 2011-06-14 at 00:40 -0700, Koen Kooi wrote: Op 13 jun 2011, om 23:30 heeft Richard Purdie het volgende geschreven: On Mon, 2011-06-13 at 22:36 +0200, Koen Kooi wrote: Op 13 jun 2011, om 22:28 heeft Saul Wold het volgende geschreven: On 06/13/2011 11:31 AM, Koen Kooi wrote: Hi, Khem was asking if I could reproduce the recent x86 breakage he was seeing[1] and I ran into another bug: koen@dominion:/OE/tentacle/sources/meta-intel$ git blame meta-n450/conf/machine/n450.conf | grep atom 158f88d7 (Saul Wold 2011-01-03 15:33:52 -0800 6) require conf/machine/atom-pc.conf meta-yocto seems to be the place you need to look! I hope that the layering tools can help to detect and inform folks of this like of dependency. Isn't meta-yocto supposed to a the integration layer with no new parts? I can't use meta-yocto since it has conflicting beagleboard stuff in it, which means that meta-intel is now broken for me as well. That surely isn't the intended plan?!?! The plan on public record is that atom-pc moves to meta-intel as soon as the layer tooling comes online and meta-yocto becomes its own repo (which at present its not but its certainly the intent). So can someone at least put that in the meta-intel or meta-n450 README? Most layers now have a README with the dependency info (e.g. http://git.shr-project.org/git/?p=meta-smartphone.git;a=blob;f=meta-palm/README;h=200a3c83ff11ad790fe4e03e93a26520ec0c6714;hb=HEAD ) It would be nice if the ones on yocto-project.org had somethign similar :) _ Would something like this do the trick for now (if so, I can do similar for the other meta-intel bsps)? I'd put it in README, but yeah, it's an improvement. ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On Tue, 2011-06-14 at 14:08 -0700, Koen Kooi wrote: Op 14 jun 2011, om 22:20 heeft Tom Zanussi het volgende geschreven: On Tue, 2011-06-14 at 00:40 -0700, Koen Kooi wrote: Op 13 jun 2011, om 23:30 heeft Richard Purdie het volgende geschreven: On Mon, 2011-06-13 at 22:36 +0200, Koen Kooi wrote: Op 13 jun 2011, om 22:28 heeft Saul Wold het volgende geschreven: On 06/13/2011 11:31 AM, Koen Kooi wrote: Hi, Khem was asking if I could reproduce the recent x86 breakage he was seeing[1] and I ran into another bug: koen@dominion:/OE/tentacle/sources/meta-intel$ git blame meta-n450/conf/machine/n450.conf | grep atom 158f88d7 (Saul Wold 2011-01-03 15:33:52 -0800 6) require conf/machine/atom-pc.conf meta-yocto seems to be the place you need to look! I hope that the layering tools can help to detect and inform folks of this like of dependency. Isn't meta-yocto supposed to a the integration layer with no new parts? I can't use meta-yocto since it has conflicting beagleboard stuff in it, which means that meta-intel is now broken for me as well. That surely isn't the intended plan?!?! The plan on public record is that atom-pc moves to meta-intel as soon as the layer tooling comes online and meta-yocto becomes its own repo (which at present its not but its certainly the intent). So can someone at least put that in the meta-intel or meta-n450 README? Most layers now have a README with the dependency info (e.g. http://git.shr-project.org/git/?p=meta-smartphone.git;a=blob;f=meta-palm/README;h=200a3c83ff11ad790fe4e03e93a26520ec0c6714;hb=HEAD ) It would be nice if the ones on yocto-project.org had somethign similar :) _ Would something like this do the trick for now (if so, I can do similar for the other meta-intel bsps)? I'd put it in README, but yeah, it's an improvement. OK, yeah, for some reason n450 decided to name its README 'ReleaseNotes'. I'll find out if there was a reason for that and rename it to README like the others if not... Tom ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On 06/13/2011 02:30 PM, Richard Purdie wrote: On Mon, 2011-06-13 at 22:36 +0200, Koen Kooi wrote: Op 13 jun 2011, om 22:28 heeft Saul Wold het volgende geschreven: On 06/13/2011 11:31 AM, Koen Kooi wrote: Hi, Khem was asking if I could reproduce the recent x86 breakage he was seeing[1] and I ran into another bug: koen@dominion:/OE/tentacle/sources/meta-intel$ git blame meta-n450/conf/machine/n450.conf | grep atom 158f88d7 (Saul Wold 2011-01-03 15:33:52 -0800 6) require conf/machine/atom-pc.conf meta-yocto seems to be the place you need to look! I hope that the layering tools can help to detect and inform folks of this like of dependency. Isn't meta-yocto supposed to a the integration layer with no new parts? I can't use meta-yocto since it has conflicting beagleboard stuff in it, which means that meta-intel is now broken for me as well. That surely isn't the intended plan?!?! The plan on public record is that atom-pc moves to meta-intel as soon as the layer tooling comes online and meta-yocto becomes its own repo (which at present its not but its certainly the intent). Until then, and even afterwards can we please get some testing of non-poky builds done? I know the autobuilder is full but can't we toss a few things onto a personal box and try that a few times a week? -- Tom Rini Mentor Graphics Corporation ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On Mon, Jun 13, 2011 at 22:10, Tom Rini tom_r...@mentor.com wrote: Until then, and even afterwards can we please get some testing of non-poky builds done? I know the autobuilder is full but can't we toss a few things onto a personal box and try that a few times a week? I think this ought to be the inverse. Poky ought to be less built since it depends (or will depends) on oe-core so this needs to be fully tested. After this passes then Poky and Yocto ought to be tested since the base has been proved stable and like. My 2c. -- Otavio Salvador O.S. Systems E-mail: ota...@ossystems.com.br http://www.ossystems.com.br Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On Mon, 2011-06-13 at 15:10 -0700, Tom Rini wrote: On 06/13/2011 02:30 PM, Richard Purdie wrote: On Mon, 2011-06-13 at 22:36 +0200, Koen Kooi wrote: Op 13 jun 2011, om 22:28 heeft Saul Wold het volgende geschreven: On 06/13/2011 11:31 AM, Koen Kooi wrote: Hi, Khem was asking if I could reproduce the recent x86 breakage he was seeing[1] and I ran into another bug: koen@dominion:/OE/tentacle/sources/meta-intel$ git blame meta-n450/conf/machine/n450.conf | grep atom 158f88d7 (Saul Wold 2011-01-03 15:33:52 -0800 6) require conf/machine/atom-pc.conf meta-yocto seems to be the place you need to look! I hope that the layering tools can help to detect and inform folks of this like of dependency. Isn't meta-yocto supposed to a the integration layer with no new parts? I can't use meta-yocto since it has conflicting beagleboard stuff in it, which means that meta-intel is now broken for me as well. That surely isn't the intended plan?!?! The plan on public record is that atom-pc moves to meta-intel as soon as the layer tooling comes online and meta-yocto becomes its own repo (which at present its not but its certainly the intent). Until then, and even afterwards can we please get some testing of non-poky builds done? I know the autobuilder is full but can't we toss a few things onto a personal box and try that a few times a week? Sure, the more people testing the various combinations the better! Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On Mon, 2011-06-13 at 15:44 -0700, Tom Rini wrote: On 06/13/2011 03:35 PM, Richard Purdie wrote: On Mon, 2011-06-13 at 15:10 -0700, Tom Rini wrote: On 06/13/2011 02:30 PM, Richard Purdie wrote: On Mon, 2011-06-13 at 22:36 +0200, Koen Kooi wrote: Op 13 jun 2011, om 22:28 heeft Saul Wold het volgende geschreven: On 06/13/2011 11:31 AM, Koen Kooi wrote: Hi, Khem was asking if I could reproduce the recent x86 breakage he was seeing[1] and I ran into another bug: koen@dominion:/OE/tentacle/sources/meta-intel$ git blame meta-n450/conf/machine/n450.conf | grep atom 158f88d7 (Saul Wold 2011-01-03 15:33:52 -0800 6) require conf/machine/atom-pc.conf meta-yocto seems to be the place you need to look! I hope that the layering tools can help to detect and inform folks of this like of dependency. Isn't meta-yocto supposed to a the integration layer with no new parts? I can't use meta-yocto since it has conflicting beagleboard stuff in it, which means that meta-intel is now broken for me as well. That surely isn't the intended plan?!?! The plan on public record is that atom-pc moves to meta-intel as soon as the layer tooling comes online and meta-yocto becomes its own repo (which at present its not but its certainly the intent). Until then, and even afterwards can we please get some testing of non-poky builds done? I know the autobuilder is full but can't we toss a few things onto a personal box and try that a few times a week? Sure, the more people testing the various combinations the better! I fear I'm not being clear. Can You, Saul and maybe other folks making frequent submissions and are at times more poky-oriented than not, do this as well? While I'd love the world I'd settle for a bunch of -g's to catch obvious problems and a console-image or something.. Since you're highlighting me personally here, I do test a variety of things periodically. I only have access to one desktop machine and one Linux laptop so just like everyone else the testing I can physically do is limited. I also merge a ton of changes from various people and rely at some level on trust of those people to have tested changes. I know Saul also does a lot of testing of various combinations. It might not always be the combination you personally want but its certainly better than no testing at all. OE is getting a number of new computer resources soon thanks to the Linux Foundation and testing OE-Core is on Tom King's todo list. Yocto is also stepping up and doing a lot of testing. It is hardware limited and also looking to increase its resources which is planned and happening, albeit slower than we'd like in an ideal world. So on the one hand I do understand your concern. I'm personally and Yocto are doing the best we can. On the other I'd suggest if testing certain combinations is this important to you (or Mentor?), stepping up and helping with the testing would be *much* appreciated and it isn't the sole responsibility of myself or Saul. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On Mon, 2011-06-13 at 23:17 +, Otavio Salvador wrote: On Mon, Jun 13, 2011 at 23:04, Richard Purdie richard.pur...@linuxfoundation.org wrote: ... So on the one hand I do understand your concern. I'm personally and Yocto are doing the best we can. On the other I'd suggest if testing certain combinations is this important to you (or Mentor?), stepping up and helping with the testing would be *much* appreciated and it isn't the sole responsibility of myself or Saul. ... It would be easier and better if people at Yocto could start basing their work on oe-core so stuff get tested there instead of Poky. Poky would then be an integration point not a base. Have you looked at the delta recently? Yocto uses OE-Core with the single addition of the meta-yocto layer which is tiny. Just like angstrom use the meta-angstrom layer and the meta-oe layer. More then once I got broken trees for stuff that were pushed to oe-core and were not working due missing fixes or features that were pushed to Poky's bitbake but not to the upstream one. Again, please look at the delta between upstream bitbake and the one in poky. All bitbake patches are now landing upstream first. There were issues, we came up with a plan to address them and we're doing what we said we would do... Doing this would help to improve it a lot. For example meta-intel would be already fixed since people would be using it against oe-core and would have already noticed the missing machine definition and like. We *know* the machine definition isn't there, its deliberate. We came up with a plan to create OE-Core and to get Poky and OE both migrated to using it. This process is not 100% complete yet although it gets closer every day. Cheers, Richard ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
Re: [OE-core] Where is atom-pc.conf hiding?
On Tue, Jun 14, 2011 at 00:02, Richard Purdie richard.pur...@linuxfoundation.org wrote: ... We *know* the machine definition isn't there, its deliberate. We came up with a plan to create OE-Core and to get Poky and OE both migrated to using it. This process is not 100% complete yet although it gets closer every day. I know the delta is getting smaller and smaller. This is good. A good way to make it move faster is using OE-Core as basis. This will force the backporting of missing Poky's change to OE-Core to happen faster. Currently there're many people that base the stuff (that is target to OE-Core) on Poky and sometimes it cases issues (as the bitbake ones I used as example). -- Otavio Salvador O.S. Systems E-mail: ota...@ossystems.com.br http://www.ossystems.com.br Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core