[PATCH] ARM: supplementing IO accessors with 64 bit capability

Mathieu Poirier mathieu.poirier at linaro.org
Wed Oct 22 10:55:55 PDT 2014


On 22 October 2014 19:19, Russell King - ARM Linux
<linux at arm.linux.org.uk> wrote:
> On Wed, Oct 22, 2014 at 06:22:09PM +0200, Mathieu Poirier wrote:
>> I had this conversation with a colleague who reviewed the work.  If
>> the architecture is < 5 the __raw_ versions aren't included and the
>> compiler won't complain until someone tries to use the macros.  We
>> achieve the same result - the macros aren't accessible when the
>> architecture doesn't support it - while saving an #if condition in the
>> file.
>>
>> I'm not strongly opinionated on this - I can enclose the macros in an
>> #if statement.
>
> What you're missing is that some driver may test for their presence by
> doing:
>
> #ifndef readq_relaxed
> ... do something else
>
> which would now break as a detection method, as the macro is always
> defined no matter whether it's present or not.

Good point - thanks for showing me the light.

>
> --
> FTTC broadband for 0.8mile line: currently at 9.5Mbps down 400kbps up
> according to speedtest.net.



More information about the linux-arm-kernel mailing list