Re: [meta-intel] [PATCH 0/5] [DORA] meta-isg: add new BSP layer for Valley Island

2014-03-21 Thread Tom Zanussi
On Fri, 2014-03-21 at 05:41 -0500, Chang, Rebecca Swee Fun wrote:
  -Original Message-
  From: Darren Hart [mailto:dvh...@linux.intel.com]
  Sent: 21 March, 2014 12:29 PM
  To: Chang, Rebecca Swee Fun; Zanussi, Tom
  Cc: meta-intel@yoctoproject.org
  Subject: Re: [meta-intel] [PATCH 0/5] [DORA] meta-isg: add new BSP layer
  for Valley Island
  
  On 3/20/14, 20:52, Chang, Rebecca Swee Fun
  rebecca.swee.fun.ch...@intel.com wrote:
  
  The changes in linux-yocto-3.10 was merged. And I sent the pull request
  for meta-intel after linux-yocto-3.10 was updated. I noticed that the
  commit ids are not the one I'm using. Is the recipe goes wrong some
  where else?
  
  I've also checked my branch (rebeccas/meta-valleyisland-dora-dev) in
  meta-intel-contrib. They are same as what I am sending out. But your
  build error log showing the wrong commit id.
  
  I need to check again to confirm.
  
  Ah, the commit Ids will be different after Bruce adds his Signed-off-by and 
  if
  he merged it with anything pending. Always check the linux-yocto tree for
  new commit Ids after your changes have been merged.
 
 No, I think my recipe is correct. I have send out a recipe with commit ids 
 with
 Bruce's signed-off. However, Tom's build failure was building at different
 commit id. 
 
 In my recipe, I am appending:
 SRCREV_machine-linux-yocto_valleyisland-32 ?= 
 61dde96f97bb5b1ed4c11caf9a857d55ad8f6e17
 SRCREV_meta-linux-yocto_valleyisland-32 ?= 
 99c503a92885060bebf2bba6747735e8e9346a40
 
 From Tom's build log:
 3.10.32+gitAUTOINC+6ad20f049a_c03195ed6e-r0
 Which does not match with my recipe.
 
 I don't have compiling issue here. Can Tom please help to check his build 
 again?
 Thanks a lot.
 

I'll check, but I don't expect any difference - this was from a clean
build after seeing it the first time.

But just to make sure we're on the same page, can you verify that you're
using the same poky and meta-intel SRCREVs as in my error message (or
send me yours)?

Thanks,

Tom

 Rebecca


-- 
___
meta-intel mailing list
meta-intel@yoctoproject.org
https://lists.yoctoproject.org/listinfo/meta-intel


Re: [meta-intel] BSP retirements from meta-intel layer

2014-03-21 Thread Chris Tapp

On 18 Mar 2014, at 01:50, Ong, Boon Leong boon.leong@intel.com wrote:

 
 
 -Original Message-
 From: Darren Hart [mailto:dvh...@linux.intel.com]
 Sent: Tuesday, March 18, 2014 6:54 AM
 To: Keskinarkaus, Teemu; Chris Tapp; Kamble, Nitin A
 Cc: meta-intel@yoctoproject.org; Chang, Rebecca Swee Fun; Ong, Boon Leong
 Subject: Re: [meta-intel] BSP retirements from meta-intel layer
 
 On 3/13/14, 21:21, Keskinarkaus, Teemu
 teemu.keskinark...@maximatecc.com wrote:
 
 From: meta-intel-boun...@yoctoproject.org [mailto:meta-intel-
 boun...@yoctoproject.org] On Behalf Of Chris Tapp
 
 Following on from the chiefriver BSP retirement thread, what is the
 general  policy for BSP life-cycles within meta-intel?
 
 For example, we committed to using the Cedartrail BSP a couple of
 years back as  if offered support for what was then a new platform
 with a reasonable forward  buy-time. However, support for that has
 been dropped even though boards are  still available using the chipset
 (e.g. ASRock DN2800MT) and these are going to  be available for a few
 years yet.
 
 From our point of view it is a pity not to be able to make use of the
 improvements  and enhancements that have gone into later Yocto
 versions, especially as we're  continually updating our software.
 
 This interests me as well since we have been planning on using Yocto on
 our HW for a long time. We also noticed the drop of Cedartrail BSP and
 we ended up on porting it to newer Yocto ourselves which of course
 wasn't what we were hoping to get when switching to Yocto.
 
 As for chief river, sys940x, and n450 (which I haven't yet posted the 
 retirement
 notification for), these systems are simply being supported by the 
 intel-common
 BSPs (intel-core2-32 and intel-corei7-64). So support is not being dropped, 
 it is
 being streamlined. The hardware is still supported.
 
 While we would like to support every platform continually, we (I speak of the
 core yocto team here, not as an Intel BSP maintainer) sometimes must decide
 between adding a new platform or continuing to support an older platform. I
 expect this to become less of an issues as the platforms become more open 
 over
 time (as we are seeing with the graphics on the Baytrail SoCs for example).
 
 As for more specific platforms such as cedar trail as mentioned here, that 
 must
 be addressed the BSP maintainers. If folks are doing the forward ports
 themselves, that is something the maintainers need to be aware of and take 
 into
 consideration with their support plans. Please discuss this with them 
 (Rebecca
 and Boon Leong added to Cc) - off list would be best as this becomes an
 individual support issue between the customer and the Board/BSP vendor (Intel
 ISG in this case).
 
 In the case of Cedartrail, the last YP BSP is support is version 1.3 (danny). 
 There is 
 no plan for upgrade to newer YP version. If you have specific need, please 
 contact
 your Intel sales rep/account and they should be able to bring your case for 
 internal
 whether an upgrade plan is approved. Thanks and sorry for inconvenience 
 caused.

Thanks for the suggestion, but I don't have an Intel sales rep/account as we 
are only a small player when it comes to hardware. We typically make 100 to 200 
units a year, and, because this is low volume, we use COTS hardware like the 
DN2800MT (formally Intel, now ASRock). I will get in touch with the UK sales 
office and see what they suggest.

In general I can't imagine anyone expecting support for old hardware to 
continue, but support was dropped with Dylan back in April 2013 even though the 
silicon is still in production. This is really why I was hoping an official 
position on support could be given, as small players (of which there are many 
in the embedded sector) don't have the contacts with the silicon vendors which 
means it can take a long, long time for important information to make it 
through.

Though there is a bit of a heads-up as I've just seen that EOL was announced 
back in November 2013 for the non-embedded version of the N2800 - just sent a 
request to our board supplier to check which one they use...

Chris Tapp

opensou...@keylevel.com
www.keylevel.com



-- 
___
meta-intel mailing list
meta-intel@yoctoproject.org
https://lists.yoctoproject.org/listinfo/meta-intel