[PATCH 2/2] MAINTAINERS: demote ARM port to "odd fixes"

[Ard Biesheuvel]

On 25 March 2018 at 19:09, Russell King - ARM Linux
<linux at armlinux.org.uk> wrote:
> On Sun, Mar 25, 2018 at 08:27:32AM +0100, Ard Biesheuvel wrote:
>> On 24 March 2018 at 15:02, Russell King <rmk+kernel at armlinux.org.uk> wrote:
>> > As of the start of 2018, I am no longer paid to support the core 32-bit
>> > ARM architecture code.  This means that this code is no longer
>> > commercially supported, and is now only supported through voluntary
>> > effort.
>> >
>>
>> Hello Russell,
>>
>> It is quite obvious what you are doing here, and FWIW, I personally
>> think you should be on the ARM payroll for looking after the 32-bit
>> port. However, I don't think it is appropriate to involve the
>> community into this. If you cannot afford to spend time on this, the
>> right thing to do would be to either share the responsibility with
>> co-maintainers (which many architecture ports do these days), or step
>> down and let someone else take over entirely.
>>
>> In my opinion, letting the ARM port wither to extort money out of ARM
>> Ltd. is not something the Linux community should be accessory to.
>
> I find your email offensive.  This is not an attempt to extort money.
> I do NOT engage in extortion.  Please apologise for that smear.
>

I did not mean to offend you. I do apologise if that is what I did.

> This merely reverts the ARM port to the situation that it was prior to
> 2001.
>
> The fact of the matter is if I'm not getting paid to look after it,
> then I'll have other paid work, and anyone who has paid work will know
> that paid work takes priority over voluntary work.  That doesn't mean I
> have to give up all interest in something I founded.
>

But how is it in the interest of the Linux community that you remain
the sole maintainer with very little time to spend on it? Maintainer
groups are working out very well in other areas, so I'd expect that to
be something you consider first before demoting the port to 'odd
fixes'.

> There are plenty of other kernel maintainers in that position - for
> example, Santosh is no longer employed by Texas, but still looks after
> the various stuff for TI.
>
> I'm merely pointing out that I'm not going to be spending quite as much
> time on the ARM port as I have done in the past - particularly, this
> means that I won't have a guaranteed amount of time each week for this
> activity.
>
> In the coming weeks, I'm going to be working on the Spectre fixes under
> contract with ARM Ltd, and as this will be my primary focus, it will
> take priority over voluntary work, such as general maintanence of the
> 32-bit ARM port.
>
> I'm sure the same goes for other people: work that you get paid for
> takes priority over voluntary work.  Anything else would be plain
> stupid and suicidal from an employment point of view.
>

Nobody is asking you to prioritise voluntary work over paid work. I am
asking you to consider offloading some of the responsibility to others
if you are failing to cope by yourself.

> The reality of the 32-bit ARM port situation is that the code base is
> stable, and it receives very few patches each kernel cycle, and so it
> is entirely reasonable that less time is spent on it.
>

If that is the case, why not simply keep it at 'maintained', which is
already documented as voluntary support rather than commercial
support.

> At no point have I said that I will not merge anything unless I'm paid
> for it - again, that would be suicidal and rediculous.
>
> What I do find particularly offensive in your attitude is the
> expectation that I'll continue to spend as much time as I have
> previously on the ARM port as I have done while being paid for it -
> that is entirely unreasonable.  I think you need to alter your
> attitude and expectations.
>

Nobody is expecting that. I merely suggested that there are other
options here besides paid full support by just you and unpaid odd
fixes by just you.

> I said in my email that I'm not going to be reading the mailing list
> quite as much because given that much of the list traffic is no longer
> relevant and sorting through it takes considerable time, I'd rather
> spend that time for more productive tasks.
>

I agree that there is a lot of noise on LAKML given that a lot of
driver patches that happen to be for IP present in ARM SoCs gets cc'd
there.