Re: [PATCH] Stop mprotect() changing MAP_SHARED and other cleanup

[Jamie Lokier]

Muli Ben-Yehuda wrote:
> > +/* Optimisation macro. */
> > +#define _calc_vm_trans(x,bit1,bit2) \
> > +  ((bit1) <= (bit2) ? ((x) & (bit1)) * ((bit2) / (bit1)) \
> > +   : ((x) & (bit1)) / ((bit1) / (bit2))
> 
> Why is this necessary? the original version of the macro was much
> simpler. If this isn't just for shaving a couple of optimization,
> please document it. If it is, I urge you to reconsider ;-) 

When the bits don't match, mine reduces to a mask-and-shift.  The
original reduces to a mask-and-conditional, which is usually slower.

Hopefully GCC optimises the latter to the former these days, but there
is no harm in helping.

I vaguely recall GCC being able to optimise (x&mask1) | (x&mask2) to
x&(mask1|mask2), not 100% sure though.  If so, the PROT_ bits
translation will reduce to prot & 7.

-- Jamie
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/