Re: [Patch 17/23] mask v2 = [6/7] nodemask_t_ia64_changes

[Paul Jackson <pj@sgi.com>]

A few hours ago, I sketched the following code:

> 	for (
> 	      ({ i = 0; while(!node_set(i, mask) && i < MAX_NUMNODES) i++; i; });
> 	      i < MAX_NUMNODES; 
> 	      ({ i++; while(!node_set(i, mask) && i < MAX_NUMNODES) i++; i; })
> 	)


I've reduced this to a pair of bitmap.h operators:

> #define find_next_bit_in_bitmap(src, nbits, n)			\
> 	({ int i = (n); while(i < (nbits) && !test_bit(i, (src))) i++; i; })
> 
> #define find_first_bit_in_bitmap(src, nbits)			\
> 	find_next_bit_in_bitmap((src), (nbits), 0)

which are plug compatible with bitops find_first_bit() and find_next_bit(),
but smaller.  These alternatives can then be used in the for-loop defines.

The above is not yet compiled or tested ...  Use at own risk.

My next mask patchset will have these.  Following the excellent
recommendations of Rusty however, my next mask patchset won't have
masks ;).  Just cpumasks bolted directly on top of souped up bitmaps.

-- 
                          I won't rest till it's the best ...
                          Programmer, Linux Scalability
                          Paul Jackson <pj@sgi.com> 1.650.933.1373