[Bug middle-end/84443] powerpc suboptimal code generation (shrink wrap unlikely path) for Linux spinlocks

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84443

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |missed-optimization
             Target|                            |powerpc
          Component|tree-optimization           |middle-end

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Note that volatiles are not optimized and thus by having

typedef struct {
 volatile unsigned int slock;
} arch_spinlock_t;

you might pessimize things.  The code looks heavily dependent on inline asms
and
inlining so it's a bit hard to analyze as well.