Re: [PATCH 1/3] mtd: nand: raw: brcmnand: Refactored code and introduced inline functions

[Boris Brezillon]

On Mon, 3 Jun 2019 10:11:20 -0400
Kamal Dasu <kdasu.kdev@xxxxxxxxx> wrote:

> Boris,
> 
> On Sat, Jun 1, 2019 at 3:57 AM Boris Brezillon
> <boris.brezillon@xxxxxxxxxxxxx> wrote:
> >
> > On Thu, 30 May 2019 17:20:35 -0400
> > Kamal Dasu <kdasu.kdev@xxxxxxxxx> wrote:
> >  
> > > Refactored NAND ECC and CMD address configuration code to use inline
> > > functions.  
> >
> > I'd expect the compiler to be smart enough to decide when inlining is
> > appropriate. Did you check that adding the inline specifier actually
> > makes a difference?  
> 
> This was done to make the code more readable. It does not make any
> difference to performance.

I meant dropping the inline specifier, not going back to manual
inlining. As a general rule, you don't need to add the 'inline'
specifier unless your function is defined in a header. In all other
cases the compiler is able to inline things on its own when it sees the
number of instructions is small enough or when the function is only
called once.