TAKE - last part of the spinlock irq change

[Steve Lord <lord@sgi.com>]

Missed one crucial part. This is a noop on i386, it is not on
plaforms where sizeof(long) != sizeof(int) and the irq mask
uses the extra bits.

This may all be a moot point if some development code pans out,
xfs will have virtually no code in interrupt context if that
works.

Steve

Date:  Fri Mar 29 08:20:57 PST 2002
Workarea:  jen.americas.sgi.com:/src/lord/xfs-linux.2.4

The following file(s) were checked into:
  bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs


Modid:  2.4.x-xfs:slinx:115179a
linux/include/linux/xfs_support/mutex.h - 1.3
	- switch mutex_spinlock to return an unsigned long