[PATCH 51/97] ARM: l2c: remove platforms/SoCs setting early BRESP

[Olof Johansson]

On Thu, May 1, 2014 at 8:12 AM, Grant Likely <grant.likely at secretlab.ca> wrote:
> On Tue, 29 Apr 2014 01:21:41 +0100, Russell King - ARM Linux <linux at arm.linux.org.uk> wrote:
>> On Tue, Apr 29, 2014 at 09:02:27AM +0900, Simon Horman wrote:
>> > On Mon, Apr 28, 2014 at 08:30:32PM +0100, Russell King wrote:
>> > > Since we now automatically enable early BRESP in core L2C-310 code when
>> > > we detect a Cortex-A9, we don't need platforms/SoCs to set this bit
>> > > explicitly.  Instead, they should seek to preserve the value of bit 30
>> > > in the auxiliary control register.
>> > >
>> > > Acked-by: Tony Lindgren <tony at atomide.com>
>> > > Signed-off-by: Russell King <rmk+kernel at arm.linux.org.uk>
>> >
>> > I would prefer if this patch was broken out into individual patches
>> > for each board or SoC file and that they were then picked up
>> > by their respective platform maintainers.
>> >
>> > Likewise for patch 66/97. Although it is only for shmobile
>> > I would prefer it broken out.
>>
>> Oh fuck that.
>>
>> Okay, I'm dropping the whole patch set right now and forgetting the whole
>> damned thing.  The L2 cache code can damned well stay as it is and remain
>> an unmaintainable mess.
>
> FWIW, there are an awful lot of people, myself included, who do care
> that you've done this work. It is 100% okay for you to say "no" to
> requests to split things up because of the complexity of the series.
>
> I really hope you're reconsider and not give up on this series.

+1. I've been very behind on email lately and I haven't paid enough
attention to this, but I'd like to see this go in for 3.16 as well,
and it shouldn't have to go in through each subtree. Sequencing that
in this case would be a huge pain indeed.

It would be better to just keep it all on a shared topic branch that
we can merge into arm-soc as a dependency, so we can resolve merge
conflicts as we merge in from platform maintainers. I wouldn't expect
anything major besides some context conflicts due to nearby edits.

The only thing I am not 100% on is if it's a good idea or not to add
to the machine descriptor, since it'll make it harder to make
descriptor-free generic platforms for A9, but it's a minor detail and
the benefit of the series outweigh that (and we can revisit if truly
needed later).


-Olof