Patch: Fix lpae on exynos5

[Peter Robinson]

On Tue, Oct 29, 2013 at 1:28 PM, Josh Boyer <jwboyer at fedoraproject.org> wrote:
> On Tue, Oct 29, 2013 at 8:48 AM, John Dulaney
> <jdulaney at fedoraproject.org> wrote:
>>> Date: Wed, 23 Oct 2013 15:53:18 +0100
>>> Subject: Re: Patch: Fix lpae on exynos5
>>> From: pbrobinson at gmail.com
>>> To: jdulaney at fedoraproject.org
>>> CC: kernel at lists.fedoraproject.org
>>>
>>> John,
>>>
>>> Is there any upstream references to this? Has this been tested against
>>> the Fedora LPAE kernel config and are there any changes needed for
>>> that, oh and which version of the Fedora kernel?
>>>
>>> Peter
>>
>> Peter,
>> The source for this patch:
>> https://github.com/virtualopensystems/linux-kvm-arm/commit/32682304c9935dd0aa55ee9196429d0955f26fa1
>>
>> I have tested it against the 3.11 and 3.12 kernels;  the file it patches is also specific to exynos5.
>
> Thanks for the follow up.  A few comments/questions.
>
> 1) If you send a patch, it needs to be properly attributed if you did
> not author it.  That includes carrying forward the From: for
> authorship and the Signed-off-by lines.
>
> 2) The tree you pointed to appears to be some random github tree.
> That is not upstream.  Upstream is the main torvalds tree, or allowing
> for ARM, the ARM maintainers tree and the ARM SoC trees.  Now, it may
> well be that this github tree is widely used and feeds into one of the
> trees I mentioned, however it is really no different than pointing to
> a random patch somewhere until those commits show up in one of those
> trees.  (Github commits don't show commentary on the patch either.)
>
> 3) The specific patch you provide was authored 6 months ago, and
> committed to github 3 months ago.  I went and looked in linux-next,
> and it isn't there.  I looked in the various branches of the arm-soc
> tree and it also isn't there.  Given the patch is half a year old now
> and it still isn't even pending in the upstream tree, we'd wind up
> carrying this as a patch basically forever.  I'm not keen on doing
> that.

The only thing I can add to this is that we've dropped the exynos
multi platform patch from 3.12 as it didn't apply cleanly and upstream
had said it would land in 3.12 which it did not. At this point in time
there is no support for exynos HW in the rawhide kernel and hence it's
not on a supported list of platforms and looking at the upstream
arm-soc branch it looks unlikely to be so until at least 3.14 so I
don't see much point in this patch landing.

Peter