Re: [PATCH 05/18] power: reset: Add AT91 reset driver

[Maxime Ripard]

On Tue, Jul 08, 2014 at 10:08:14AM +0200, Maxime Ripard wrote:
> On Mon, Jul 07, 2014 at 08:40:01PM +0200, Jean-Christophe PLAGNIOL-VILLARD wrote:
> > > > > >> move this to an assembly file more easy to read than a C code
> > > > > > 
> > > > > > Nope. It's a pain to pass variable to an external assembly file, and
> > > > > > this makes the use of global variable pretty much mandatory, which is
> > > > > > definitely not great.
> > > > > 
> > > > > Not at all I did for the PM slow clock code just write a function and pas it as a parameter
> > > > > you have 3
> > > > > 
> > > > > so basically you have to use the current and just pass at91_ramc_base[0], at91_ramc_base[1]
> > > > > and at91_rstc_base
> > > > > itâs 3 lignes of modification, if you have at91_ramc_base and at91_ramc_base same
> > > > > 
> > > > 
> > > > Yes, retrieving function parameters from assembly code is not that
> > > > complicated (the first 4 pointer values are accessible through r0-r3),
> > > > but then you'll have to store your assembly file somewhere.
> > > 
> > > Like I was saying, there's a strong preference for the inline
> > > assembly...
> > 
> > inline is horrible to read and maintain NACK
> > 
> > keep it in an assembly file it's so easy to read and follow
> > 
> > and you just have to move the file existing to the driver/power
> 
> Well, the whole rest of the kernel community feels otherwise.

Thinking a bit more about this, would using symbolic names instead of
the indices in the inline assembly work for you?

Maxime

-- 
Maxime Ripard, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

Attachment: signature.asc
Description: Digital signature