Re: Linux 2.4.10-pre11 -- __builtin_expect

[Andreas Schwab <schwab@suse.de>]

James Antill <james@and.org> writes:

|> Andi Kleen <ak@suse.de> writes:
|> 
|> > Good point. I somehow assumed that __builtin_expect would just signify
|> > a boolean, but if I read gcc source correctly this was wrong.
|> 
|>  Yeh, it's a long so you'll get no cast warnings too.
|> 
|> > Here is an updated patch.
|> 
|> [snip ... ]
|> 
|> > --- include/linux/kernel.h-LIKELY	Tue Sep 18 11:12:20 2001
|> > +++ include/linux/kernel.h	Tue Sep 18 14:35:17 2001
|> > @@ -171,4 +171,14 @@
|> >  	char _f[20-2*sizeof(long)-sizeof(int)];	/* Padding: libc5 uses this.. */
|> >  };
|> >  
|> > +
|> > +/* This loses on a few early 2.96 snapshots, but hopefully nobody uses them anymore. */ 
|> > +#if __GNUC__ > 2 || (__GNUC__ == 2 && _GNUC_MINOR__ == 96)
|> > +#define likely(x)  __builtin_expect(!!(x), 1) 
|> > +#define unlikely(x)  __builtin_expect((x), 0) 
|> 
|>  unlikely() also needs to be...
|> 
|> #define unlikely(x)  __builtin_expect(!(x), 1) 
|> 
|> ...or...
|> 
|> #define unlikely(x)  __builtin_expect(!!(x), 0) 

This is not needed, since only 0 is the likely value and !! does not
change that.

Andreas.

-- 
Andreas Schwab                                  "And now for something
Andreas.Schwab@suse.de				completely different."
SuSE Labs, SuSE GmbH, Schanzäckerstr. 10, D-90443 Nürnberg
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5