[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <31bfcd7b-246e-50b9-081f-80ba704b4e64@gmail.com>
Date: Thu, 17 Aug 2017 11:41:57 -0700
From: Florian Fainelli <f.fainelli@...il.com>
To: Eric Anholt <eric@...olt.net>,
Florian Fainelli <f.fainelli@...il.com>
Cc: linux-rpi-kernel@...ts.infradead.org,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
Stefan Wahren <stefan.wahren@...e.com>,
bcm-kernel-feedback-list@...adcom.com
Subject: Re: [GIT PULL 2/2] bcm2835-soc-next-2017-08-15
On 08/17/2017 09:12 AM, Eric Anholt wrote:
> Florian Fainelli <f.fainelli@...il.com> writes:
>
>> On 08/15/2017 11:03 AM, Eric Anholt wrote:
>>> The following changes since commit f29c256853b7412961d3ee80ca525bd2530573db:
>>>
>>> ARM: dts: bcm283x: Add 32-bit enable method for SMP (2017-08-14 20:09:44 +0200)
>>>
>>> are available in the git repository at:
>>>
>>> git://github.com/anholt/linux tags/bcm2835-soc-next-2017-08-15
>>>
>>> for you to fetch changes up to 067b437e55a892e3ebb13e40c98825fcfa1e2d99:
>>>
>>> ARM: bcm2836: Send event when onlining other cores (2017-08-15 10:52:26 -0700)
>>>
>>> ----------------------------------------------------------------
>>> This pull request brings in two things.
>>>
>>> One is to use sev() to wake up CPUs that might be sleeping when doing
>>> the custom spin-table boot process in 32-bit mode (new firmware
>>> versions will have the CPUs sleeping waiting for an event instead of
>>> just spinning). However, the irqchip maintainer objected to our SMP
>>> code continuing to live in the driver, so we had to move it to
>>> platsmp.c, and to do that we needed a new SMP enable-method to the DT
>>> for the platsmp.c to attach to (thus the DT cross-merge in this PR).
>>> The platsmp.c patch was acked by irqchip for going through arm-soc.
>>
>> This does make us pull quite a lot of changes, how about I just
>> cherry-pick "ARM: dts: bcm283x: Add 32-bit enable method for SMP" such
>> that the branch in itself is functional as-is, but we don't pull in
>> everything else from devicetree/next?
>
> Then you get the commit duplicated in the history, which people
> generally dislike even more. Also, it depends on the arm64->arm move,
> so you'd need that as well.
Alright, merged and now submitted, thanks! --
--
Florian
Powered by blists - more mailing lists