Re: [cpumask_t 1/3] core changes for 2.5.67-bk6

Randy.Dunlap (rddunlap@osdl.org)
Fri, 18 Apr 2003 10:20:15 -0700


On Tue, 15 Apr 2003 15:50:36 -0700 William Lee Irwin III <wli@holomorphy.com> wrote:

| Core changes for extended cpu masks. Basically use a machine word
| #if NR_CPUS < BITS_PER_LONG, otherwise, use a structure with an array
| of unsigned longs for it. Sprinkle it around the scheduler and a few
| other odd places that play with the cpu bitmasks. Back-ended by a
| bitmap ADT capable of dealing with arbitrary-width bitmaps, with the
| obvious micro-optimizations for NR_CPUS < BITS_PER_LONG and UP.
|
| NR_CPUS % BITS_PER_LONG != 0 is invalid while NR_CPUS > BITS_PER_LONG.

Where/why this restriction (above)?
I don't see the need for it or implementation of it.

I'm only looking at the core patch.

| diff -urpN linux-2.5.67-bk6/include/linux/bitmap.h cpu-2.5.67-bk6-1/include/linux/bitmap.h
| --- linux-2.5.67-bk6/include/linux/bitmap.h 1969-12-31 16:00:00.000000000 -0800
| +++ cpu-2.5.67-bk6-1/include/linux/bitmap.h 2003-04-15 14:39:40.000000000 -0700

| +static inline void bitmap_shift_left(volatile unsigned long *,volatile unsigned long *,int,int);

Do you need this prototype? I don't see why.

Rest of core looks good to me.

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