Re: Hardware pack questions

2010-09-14 Thread Steve Langasek
On Tue, Sep 14, 2010 at 09:16:15AM +0200, Alexander Sack wrote:

  Is it intended that pre-release hwpacks will be long-lived?  I expected
  the same rules to apply as to pre-release images:  ephemeral objects
  used during development that would be replaced at release time by images
  built from the final archive, leaving no lingering obligation to provide
  easy access to source for images we're no longer distributing.   Are the
  hwpacks not going to follow this same release cycle?

 Right, I think legally we might really not be required to do this;
 however, technically it still makes sense to me to have the sources
 used to produce a certain hwpack. Especially since we use ppa's or
 some other archives for hwpacks which have no guarantees to keep
 history etc.

Yes, if we're going to be building hwpacks out of ppas, getting the sources
out of the PPA afterwards would be a problem; so best to keep them with the
hwpack.

-- 
Steve Langasek   Give me a lever long enough and a Free OS
Debian Developer   to set it on, and I can move the world.
Ubuntu Developerhttp://www.debian.org/
slanga...@ubuntu.com vor...@debian.org


signature.asc
Description: Digital signature
___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-09-13 Thread Scott Bambrough
On Mon, 2010-09-13 at 10:50 +0200, Alexander Sack wrote:
 If the separate lexbuilder backend deployment causes any issues we can
 also just hook this up to live-helper so we produce the hwpacks in the
 headless run.
 
No, this shouldn't be the case.  Do it right the first time, and have a
longer term vision in mind.  A hack like Alexander proposes is just a
waste of time and effort IMHO.

Scott






___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-09-13 Thread Alexander Sack
On Mon, Sep 13, 2010 at 11:49 AM, Scott Bambrough
scott.bambro...@linaro.org wrote:
 On Mon, 2010-09-13 at 10:50 +0200, Alexander Sack wrote:
 If the separate lexbuilder backend deployment causes any issues we can
 also just hook this up to live-helper so we produce the hwpacks in the
 headless run.

 No, this shouldn't be the case.  Do it right the first time, and have a
 longer term vision in mind.  A hack like Alexander proposes is just a
 waste of time and effort IMHO.

If making a standalone hwpack backend cause delay to linaro-n cycle,
the effort to hook it up in live-helper is really worth it imo. It's
not much effort after all to put it in a lh helper script, so the
waste (if you really want to call it that way) would be well confined.

-- 

 - Alexander

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-09-13 Thread Scott Bambrough
On Mon, 2010-09-13 at 11:56 +0200, Alexander Sack wrote:
 On Mon, Sep 13, 2010 at 11:49 AM, Scott Bambrough
 scott.bambro...@linaro.org wrote:
  On Mon, 2010-09-13 at 10:50 +0200, Alexander Sack wrote:
  If the separate lexbuilder backend deployment causes any issues we can
  also just hook this up to live-helper so we produce the hwpacks in the
  headless run.
 
  No, this shouldn't be the case.  Do it right the first time, and have a
  longer term vision in mind.  A hack like Alexander proposes is just a
  waste of time and effort IMHO.
 
 If making a standalone hwpack backend cause delay to linaro-n cycle,
 the effort to hook it up in live-helper is really worth it imo. It's
 not much effort after all to put it in a lh helper script, so the
 waste (if you really want to call it that way) would be well confined.
 
No, James and I should be able to solve the deployment problem.  It
would be better to concentrate on the actual hardware packs themselves
and fix/improve the backend to meet our needs.  The infrastructure to
create and deploy an individual hardware pack now exists and needs real
world testing.  In order to test and shake out problems with the backend
we really need several hardware packs defined.

James can help out with the mechanics of creating the hardware pack and
getting them built automatically.  However the contents of the hardware
pack should be driven by those in Linaro defining the images we want to
build.  I haven't seen a relevant head to use with the hardware packs
yet, and that IMHO is not an infrastructure problem.  What head is
available for combination with a hardware pack by linaro-media-create is
available?

Scott



___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-09-13 Thread Alexander Sack
On Mon, Sep 13, 2010 at 12:25 PM, Scott Bambrough
scott.bambro...@linaro.org wrote:
 On Mon, 2010-09-13 at 11:56 +0200, Alexander Sack wrote:
 On Mon, Sep 13, 2010 at 11:49 AM, Scott Bambrough
 scott.bambro...@linaro.org wrote:
  On Mon, 2010-09-13 at 10:50 +0200, Alexander Sack wrote:
  If the separate lexbuilder backend deployment causes any issues we can
  also just hook this up to live-helper so we produce the hwpacks in the
  headless run.
 
  No, this shouldn't be the case.  Do it right the first time, and have a
  longer term vision in mind.  A hack like Alexander proposes is just a
  waste of time and effort IMHO.

 If making a standalone hwpack backend cause delay to linaro-n cycle,
 the effort to hook it up in live-helper is really worth it imo. It's
 not much effort after all to put it in a lh helper script, so the
 waste (if you really want to call it that way) would be well confined.

 No, James and I should be able to solve the deployment problem.  It
 would be better to concentrate on the actual hardware packs themselves
 and fix/improve the backend to meet our needs.  The infrastructure to
 create and deploy an individual hardware pack now exists and needs real
 world testing.  In order to test and shake out problems with the backend
 we really need several hardware packs defined.

Right, I sent the info about initial hwpack content in a previous mail.


 James can help out with the mechanics of creating the hardware pack and
 getting them built automatically.  However the contents of the hardware
 pack should be driven by those in Linaro defining the images we want to
 build.  I haven't seen a relevant head to use with the hardware packs
 yet, and that IMHO is not an infrastructure problem.  What head is
 available for combination with a hardware pack by linaro-media-create is
 available?

headless is already available and can already be used for testing
(even if omap3 is in there atm). The other heads are failing to build
because we don't have hardware packs yet.

Now that you say that we won't hook up hwpack-create there, i will
make a new live-helper that allows runs without kernel etc. and then
we should be able to get the alip/efl/plasma heads without hw support
rather quickly.

James, please ping me on IRC to get info how to test these with headless.

-- 

 - Alexander

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-09-13 Thread Alexander Sack
On Mon, Sep 13, 2010 at 4:03 PM, James Westby
james.wes...@canonical.com wrote:
 On Mon, 13 Sep 2010 10:50:29 +0200, Alexander Sack a...@linaro.org wrote:
 You can see the hwpacks at

  http://jameswestby.net:8080/view/Hardware%20Packs/

awesome ... i will check those a bit later.


 Unfortunately the names change, and I don't know how to set up a
 current link. However, there are stable URLs that get you most of the
 way there, e.g.

  http://jameswestby.net:8080/view/Hardware%20Packs/job/linaro-omap3%20hwpack/lastSuccessfulBuild/artifact/

I know that some partners might have problems to access this on port
8080 ... is there a way to make these available on port 80 for now?


 I can give others access to trigger builds etc. on request, and can also
 add people to the list of recipients of the build failure emails.

I think Steve L., Jamie (and me for a while) probably want those mails.



 Also, the Name has the same restrictions as the package name, so I
 just took the linaro-omap3 style, rather than the Linaro OMAP3 one.

OK. thought it was the readable name field. fine with me.


 linaro-bsp-ux500:
  + Name: Linaro BSP UX500
  + Architectures: armel
  + Origin: Linaro (BSP)
  + Maintainer: Linaro BSP Team linaro-dev@lists.linaro.org
  + Support: unsupported
  + Archives: Ubuntu main/universe, linaro overlay, linaro kernel ppa
  + with dep Packages: linux-image-ux500
  + without dep: u-boot-linaro-ux500 xserver-xorg-video-fbdev

 http://jameswestby.net:8080/view/Hardware%20Packs/job/linaro-bsp-ux500/lastBuild/console

 The cache has no package named 'linux-image-ux500'

yes, jcribgy did not add the meta package to the kernel ppa yet. CCed
him so we can have those asap.

  2. if vendors can ship non-free, click-through hwpacks for highly
 proprietary graphics, codec and other support

 What do you mean by click-through?

 This wasn't highlighted in the spec, so if it is important there will be
 a little more work to support this.

This is outside of the spec. the idea is that if hwpacks can only be
distributed through special means (aka click through, registration),
vendors can put those full packs on their own website with their own
click through mechanism etc. In this way we don't need to bother about
the problems with non-free packs.

So for now, we as linaro won't distribute hwpacks in that way in a
daily fashion.

  2. update linaro-image-tools package to contain your latest goodies

 I'm awaiting some further reviews from my team in order to land all of
 the branches that are being used to create them.

OK, let me know if you need anything from our side.


  4. add support for installing hwpacks in linaro-media-create

 Salgado was working on this, but has been out for a few days. It's
 obviously important to be able to install.

 Guilherme, are you happy to continue working on this, or would you like
 me to carry on your work?

  5. either add linaro-image-tools to headless and head images
  5a. ... or install and remove it during linaro-media-creation
  6. update plars download image script to get the appropriate hwpack
 together with the head the user wants.

 How do we work hwpacks in to the ISO tracker?


I have no strong opinion on how we do that. I think Paul had some ideas (CCed)

 Thanks,


Thanks to you for your great work!

-- 

 - Alexander

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-09-13 Thread Alexander Sack
On Mon, Sep 13, 2010 at 4:03 PM, James Westby
james.wes...@canonical.com wrote:
 On Mon, 13 Sep 2010 12:35:05 +0200, Alexander Sack a...@linaro.org wrote:
 headless is already available and can already be used for testing
 (even if omap3 is in there atm). The other heads are failing to build
 because we don't have hardware packs yet.

 Now that you say that we won't hook up hwpack-create there, i will
 make a new live-helper that allows runs without kernel etc. and then
 we should be able to get the alip/efl/plasma heads without hw support
 rather quickly.

 What do you mean runs without kernel etc.?

live-helper refuses to run if you don't add any linux_flavour. its a
simple fix and i already did that in my ppa live-helper:

From changelog:
* functions/defaults.sh:
+ in turn allow builds without kernel flavours

https://edge.launchpad.net/~asac/+archive/armel1/+files/live-helper_2.0%7Ea10.1-1ubuntu1%7Elinaro3.dsc


-- 

 - Alexander

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-09-13 Thread James Westby
On Mon, 13 Sep 2010 17:03:16 +0200, Alexander Sack a...@linaro.org wrote:
 Sure: with this change we can produce our images without any kernel
 whatsoever (e.g. as those are coming from hwpacks).

Ok, we're back to this again. I still haven't heard any convincing
arguments for why that is a good idea. Could you please tell me why you
think we should do that?

Thanks,

James

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-09-03 Thread James Westby
On Fri, 27 Aug 2010 15:13:11 -0300, Christian Robottom Reis k...@linaro.org 
wrote:
 Yes. Scott B. or Ian may have a linaro-infrastructure project or project
 group in the wings to which we should move it later if so, but don't let
 yourself get blocked for lack of a place to put it ;-)

Created:

  https://blueprints.launchpad.net/linaro/+spec/infrastructure-m-hardware-packs

Does anything need to be done such that it will get picked up by the
tracker etc?

As for the spec itself, it's rather light on the endgame right now. What
does success look like for this project? What further steps do we need
to take once we have scripts to create and install hardware packs, and
the ability to generate them in lexbuilder?

Already a workitem: Branch(es) containing the hardware pack
specifications for hardware packs we want to contain daily.

Other ideas: 

  * documentation: of what? where?
  * distribution: does the existing system we have for images cover us
   here?
  * testing: do we need to add these to the ISO tracker?

What criteria are we going to use to judge if it was worth our time to
do this work? How are we going to decide if we are better off with
hardware packs, or more images?


Thanks,

James

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-27 Thread James Westby

[ resending with the correct address ]

On Fri, 27 Aug 2010 14:03:32 -0400, James Westby james.wes...@canonical.com 
wrote:
 On Fri, 20 Aug 2010 16:26:46 -0400, James Westby james.wes...@linaro.org 
 wrote:
  There is also one larger question, which is that I disagree that we
  shouldn't provide anything that will go in a hardware pack in the linaro
  images. I think that having the images be useful by themselves has lots
  of benefits.
  
- Being able to quickly spin up an image in QEMU makes for easier
  testing.
- Requiring a hardware pack for every operation will make some things
  more tedious.
- If everything in a hardware pack becomes more consolidated then
  hardware packs probably become less useful.
 - Not having a flag day where suddenly our images aren't installable
 and hwpack-install has to work well, and before that date we can't
 test hwpack-install on the images we produce.
 
 Having not had anyone convince me that stripping our images is the right
 thing to do I have carried on attempting to write the spec without
 requiring this. There will be a few issues, such as ensuring that the
 kernel we want is the one that boots, but we have that problem on hwpack
 upgrades anyway, so it doesn't go away by stripping the images.
 
 I have
 
   https://wiki.linaro.org/Platform/UserPlatforms/Specs/10.11/HardwarePacks
 
 to a state where I am happy to start implementation now. Feedback on the
 spec is still welcome, and things will still be subject to change. In
 particular there are still a number of TODOs in the spec where I don't
 know how to proceed, but I believe that none of those block us starting
 implementation of other parts.
 
 Is the current status quo to create specs under the linaro project on
 Launchpad? I'll create a spec for this so that we can track work items
 for it.
 
 Thanks,
 
 James
 

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-27 Thread Christian Robottom Reis
On Fri, Aug 27, 2010 at 02:05:30PM -0400, James Westby wrote:
https://wiki.linaro.org/Platform/UserPlatforms/Specs/10.11/HardwarePacks
  
  to a state where I am happy to start implementation now. Feedback on the
  spec is still welcome, and things will still be subject to change. In
  particular there are still a number of TODOs in the spec where I don't
  know how to proceed, but I believe that none of those block us starting
  implementation of other parts.
  
  Is the current status quo to create specs under the linaro project on
  Launchpad? I'll create a spec for this so that we can track work items
  for it.

Yes. Scott B. or Ian may have a linaro-infrastructure project or project
group in the wings to which we should move it later if so, but don't let
yourself get blocked for lack of a place to put it ;-)

Thanks very much for the help with this, James.
-- 
Christian Robottom Reis   | [+55] 16 9112 6430 | http://launchpad.net/~kiko
Linaro Engineering VP | [ +1] 612 216 4935 | http://async.com.br/~kiko

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-27 Thread Jamie Bennett
On 27 Aug 2010, at 19:05, James Westby wrote:
 On Fri, 27 Aug 2010 14:03:32 -0400, James Westby james.wes...@canonical.com 
 wrote:
 On Fri, 20 Aug 2010 16:26:46 -0400, James Westby james.wes...@linaro.org 
 wrote:
 Is the current status quo to create specs under the linaro project on
 Launchpad? I'll create a spec for this so that we can track work items
 for it.

Yes. 

The Ubuntu project was chosen for the initial blueprints for a number of 
reasons, all of which were not really valid (we had a lot of discussions about 
this) but from now on, unless your blueprint is a working group blueprint, 
please target at the Linaro project.

 Thanks,
 
 James

Regards,
Jamie.
___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-24 Thread Scott Bambrough
On Mon, 2010-08-23 at 17:06 +0200, Alexander Sack wrote:
  2. What is the purpose of the hwpack.deb that is mentioned in
 places?
 
 this is scotts baby i think. personally i am fine without a
 hwpack.deb. I think the idea was that configs etc. like apt source
 lines accompanying the hwpack could be shipped in there. Personally I
 think its good enough to start with this meta info being optionally in
 the tarballs somewhere.

This was my idea.  It was to put everything about the hardware pack in a
deb, and install it first.  On second thought, it seems overkill, and it
should probably be scrubbed from the spec. 

Scott

-- 
Scott Bambrough scott.bambro...@linaro.org
Linaro Infrastructure Team Lead


___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-23 Thread Alexander Sack
On Mon, Aug 23, 2010 at 5:14 PM, James Westby james.wes...@linaro.orgwrote:

 On Mon, 23 Aug 2010 17:06:18 +0200, Alexander Sack a...@linaro.org
 wrote:
  in theory yes, but practically I don't expect this to become a major use
  case. If it's easier assuming that there is just one in the first phase,
  lets do that.

 The big problem I see is knowing which hwpacks should supercede all
 others. Doing it by name makes sense for upgrades, but there could be
 many that could install a kernel.


Yeah. hwpacks should have a unique id in their meta info. In that way you
can figure this.



 Maybe this doesn't actually matter, but my suspicion is that it will.

  this is scotts baby i think. personally i am fine without a hwpack.deb. I
  think the idea was that configs etc. like apt source lines accompanying
 the
  hwpack could be shipped in there. Personally I think its good enough to
  start with this meta info being optionally in the tarballs somewhere.

 I think that's fine. The spec is just confusing as it suddenly starts
 discussing it in the middle.


good. whatever is easiest.



  we dont want to have magic that pulls a new hwpack at this stage.
  hwpack-install is used to install and upgrade from a local/remote-url
 that
  points to the hwpack tarball.
 
  However, as (iirc) mentioned in the spec, a hardware pack optionally can
  also ship custom apt lines, so magic upgrades without a new hwpack
 tarball
  can happen for hwpacks that come with such a location (e.g. a ppa or a
  partner hosted apt repository etc.).

 That's upgrades of the packages, not upgrades of the hwpack.


 exactly.


5. What are the use cases for tags? I can only see X/no X in the spec.
  
 
  tags are low priority. It came up with an ability to not install parts of
  the pack (like no x drivers if you dont care about X in a headless
 install
  for instance.

 Ok, tags are deferred from the initial implementation. Once we have some
 experience using them we will be able to define the user interaction
 much better.


Thanks. Thats fine.




6. What are the use cases for support information?
  
  
  We want to label hwpacks as unsupported or community so we can offer
  them to download on snapshots.linaro.org while sending a clear message
 that
  those are not official linaro hardware packs because they don't come from
 a
  linaro consolidated kernel tree for instance.

 Does that information have to go inside the hwpack. What would display
 that information? What would behave differently?


The canonical meta info like this should go into the hwpack itself. On top
we would also want to have that in the hwpack filename, so its obvious
without introspection in the download.


-- 

 - Alexander
___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-23 Thread James Westby
On Mon, 23 Aug 2010 19:52:47 +0200, Alexander Sack a...@linaro.org wrote:
 I thought a bit more about this and i think the single hwpacks policy makes
 the clean up part mentioned in last sentence of user story 2 easier to
 implement.

Right.

Maybe we want hardware pack types in the future, so that you can have 1
of the kernel type, one of the graphics type, etc.

This is where I see some blurring of the lines between a hwpack and the
packages that it contains though, so something feels wrong about that
strategy.

I will update the spec to indicate that only a single hwpack can be
installed at one time.

Thanks,

James

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-23 Thread Alexander Sack
On Mon, Aug 23, 2010 at 8:29 PM, James Westby james.wes...@linaro.orgwrote:

 On Mon, 23 Aug 2010 17:06:18 +0200, Alexander Sack a...@linaro.org
 wrote:
6. What are the use cases for support information?
  
  
  We want to label hwpacks as unsupported or community so we can offer
  them to download on snapshots.linaro.org while sending a clear message
 that
  those are not official linaro hardware packs because they don't come from
 a
  linaro consolidated kernel tree for instance.

 How is this different from the support status of the packages that the
 hardware pack installs? Should it be derived from that data.


Do we already have a linaro support status for packages implemented? or are
you refering to the ubuntu style main/universe/package-set support status
here?

-- 

 - Alexander
___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-23 Thread James Westby
On Mon, 23 Aug 2010 20:31:37 +0200, Alexander Sack a...@linaro.org wrote:
 Do we already have a linaro support status for packages implemented? or are
 you refering to the ubuntu style main/universe/package-set support status
 here?

I'm asking both conceptually and concretely.

In theory, how does the support status differ from the support status of
the package? I assume that we wouldn't have a supported hardware pack
based on an unsupported kernel, but would we ever want to release an
unsupported hardware pack containing a supported kernel?

Concretely if we have rules then this could be based on the Supported
metadata in the package indices. This is probably not available in PPAs
etc., so may not work, but let's decide if it's even something we want
to do first.

Thanks,

James

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev


Re: Hardware pack questions

2010-08-21 Thread James Westby
On Fri, 20 Aug 2010 19:15:31 -0300, Christian Robottom Reis 
k...@canonical.com wrote:
 I'm not sure I understand your question, though. Are you asking if
 packages could be excluded at hardware-pack install time or at creation
 time?

I mean at install time.

The only use case I have seen so far is --no-X. That's fine to support,
we can just mark whether a package is X related in some way, and so
exclude it if necessary.

I would like to know of any other use cases though.

 One thing which is nice about tags is that it's clear what platforms are
 supported in which configurations. So you can look at board X and see
 ah, it has a no-X gstreamer hardware pack, but no X gstreamer pack,
 which means at the moment there's no X acceleration for it.

Using X to mean x.org and an arbitrary board is mighty confusing :-)

Your use case above implies some more tools than are currently in the
spec. How much nicer is it than just seeing which packages are
installed?

Thanks,

James

___
linaro-dev mailing list
linaro-dev@lists.linaro.org
http://lists.linaro.org/mailman/listinfo/linaro-dev