PIII SIMD support in gas

[Alan Modra]

Hi Doug,
   Recently, I checked in a slightly modified version of your PIII SIMD
patches to the cvs binutils sources.  On your last email, you said you had
access to PIII systems, and could verify correct coding for some
instructions.  Is that still possible?  I'm particularly interested in
finding out whether movmskps and pinsrw are generating the correct modrm
bytes.  Your original patch disagreed with Intel documentation on 
movmskps (I wouldn't be surprised if the doco is wrong, although I
followed it), and the pinsrw encoding isn't self-consistent.

If anyone else would like to help with this, make sure that %cr4 bit 9
is set, %cr0 bit 2 is reset, set %eax and %ecx to -1, then execute the
following opcode on a PIII under gdb, and tell me whether %eax or %ecx is
modified.

0x0f,0x50,0xc1		# movmskps

Testing pinsrw is a little more difficult, and I'm not so concerned about
it.