and everyone...
On 22 September 2011 11:48, Zach Pfeffer zach.pfeffer@linaro.org wrote:
Adding kiko, Alexander and Scott.
On 22 September 2011 10:39, Zach Pfeffer zach.pfeffer@linaro.org wrote:
I just wanted to get a thread going on this.
Andy's idea:
"What I think we need is for Linaro to take charge of common-3.x, ie, a linaro-common-tracking branch. That will let anyone 'androidize' Linus HEAD based trees, which you can't do routinely at the moment. I guess usually, it's no so painful to continuously uplevel Android patchset day by day unless some gross conflict is coming."
John's opinion is that this upleveling isn't trivial and that we should wait for Google to do it since they're the experts.
I also share this opinion. I don't think we should diverg from Google's official kernel/common.git by pre-uplevaling since the result won't be a valid Android tree.
The problem is, it would be useful to have a tracking Android branch so that people who use topic trees can more easily octopus merge them together to produce an Android tree, the alternative is to cherry-pick patch sets onto jstultz's from trees who have a different basis.
Usually Android's kernel/common.git is upreved more often, but with kernel.org being down and http://android.git.kernel.org/ we're currently in a weird situation.
So the question is,
Its non-trivial to take a LT or WG tree, which is based on the latest kernel and move it into a platform, as the number of platforms increases this will become a bigger job. We've managed to get by, but we should come up with a cohesive plan so that we can meet this challenge. Thoughts?
-- Zach Pfeffer Android Platform Team Lead, Linaro Platform Teams Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#%21/linaroorg - http://www.linaro.org/linaro-blog
-- Zach Pfeffer Android Platform Team Lead, Linaro Platform Teams Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#%21/linaroorg - http://www.linaro.org/linaro-blog
On 22 September 2011 11:50, Zach Pfeffer zach.pfeffer@linaro.org wrote:
and everyone...
On 22 September 2011 11:48, Zach Pfeffer zach.pfeffer@linaro.org wrote:
Adding kiko, Alexander and Scott.
On 22 September 2011 10:39, Zach Pfeffer zach.pfeffer@linaro.org wrote:
I just wanted to get a thread going on this.
Andy's idea:
"What I think we need is for Linaro to take charge of common-3.x, ie, a linaro-common-tracking branch. That will let anyone 'androidize' Linus HEAD based trees, which you can't do routinely at the moment. I guess usually, it's no so painful to continuously uplevel Android patchset day by day unless some gross conflict is coming."
John's opinion is that this upleveling isn't trivial and that we should wait for Google to do it since they're the experts.
I also share this opinion. I don't think we should diverg from Google's official kernel/common.git by pre-uplevaling since the result won't be a valid Android tree.
The problem is, it would be useful to have a tracking Android branch so that people who use topic trees can more easily octopus merge them together to produce an Android tree, the alternative is to cherry-pick patch sets onto jstultz's from trees who have a different basis.
Usually Android's kernel/common.git is upreved more often, but with kernel.org being down and http://android.git.kernel.org/ we're currently in a weird situation.
So the question is,
Its non-trivial to take a LT or WG tree, which is based on the latest kernel and move it into a platform, as the number of platforms increases this will become a bigger job. We've managed to get by, but we should come up with a cohesive plan so that we can meet this challenge. Thoughts?
Further,
Should we just shoot for only making Android builds from upstreamed work (kernel.org and AOSP) and not worry about hardware enablement unless that enablement can be done in an upstreamable way?
-- Zach Pfeffer Android Platform Team Lead, Linaro Platform Teams Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#%21/linaroorg - http://www.linaro.org/linaro-blog
-- Zach Pfeffer Android Platform Team Lead, Linaro Platform Teams Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#%21/linaroorg - http://www.linaro.org/linaro-blog
-- Zach Pfeffer Android Platform Team Lead, Linaro Platform Teams Linaro.org | Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro http://twitter.com/#%21/linaroorg - http://www.linaro.org/linaro-blog
On Thu, Sep 22, 2011 at 04:24:17PM -0500, Zach Pfeffer wrote:
"What I think we need is for Linaro to take charge of common-3.x, ie, a linaro-common-tracking branch. That will let anyone 'androidize' Linus HEAD based trees, which you can't do routinely at the moment. I guess usually, it's no so painful to continuously uplevel Android patchset day by day unless some gross conflict is coming."
John's opinion is that this upleveling isn't trivial and that we should wait for Google to do it since they're the experts.
So how do we accomplish this common tracking branch across new releases of upstream? Or am I misunderstanding Andy's idea -- is he not suggesting we have something which is Linus HEAD plus Android patches?
Should we just shoot for only making Android builds from upstreamed work (kernel.org and AOSP) and not worry about hardware enablement unless that enablement can be done in an upstreamable way?
If the question is simply that, the answer is of course no -- the Landing Team trees can contain stuff which is not-yet-upstreamable, and we should still use them for Android builds if they can be used for that.
But I think you're asking a different question.