svn commit: r232351 - in head/sys: kern sys ufs/ffs ufs/ufs

[David Schultz]

On Fri, Mar 02, 2012, David Chisnall wrote:
> On 2 Mar 2012, at 12:53, Konstantin Belousov wrote:
> 
> > This part of the change breaks KBI. I suggest that for merge to stable/9 
> > you would leave the bread and breadn as functions.
> 
> Can we not do this for the general case?  Provide them as inline
> extern functions in the header, and implement them elsewhere, so
> the compiler will inline them in recompiled code but not break
> code that isn't?

The biggest hinderance to using extern inline is that gcc and C99
disagree about what it means, unless you use a reasonably recent
compiler in C99 mode.  I first tried to use extern inline in the
tree several years after I backported gcc's C99 inline support,
and it still turned out to be a headache.

Of course, the kernel is easier because we can insist on a recent
compiler in C99 mode, but perhaps there's still some lingering
confusion.  I've been meaning to write up a wiki page about how to
use the different types of inlines, and when they'd be
appropriate...but I have very limited time for the next few months.