NPX init bugs in x86 head.S ???

[<cutaway@bellsouth.net>]

Below is the code snip that checks for a hardware NPX - which it does OK.

But, how can enabling a 287 ever work in a paged environment?  A 287 NPX
only has internal registers capable of handling 16:16 segmented protected
mode addresses.  Once paging is enabled and 16:32 addresses start getting
generated I should think everything falls apart.  At a minimum, if nothing
else died first, any reported exception CS:IP's would certainly be bogus
having only 16 bits of IP coming back to the 386 from the NPX.

Am I missing something here?

 Tony

------------------CUT-------------------

/*
 * We depend on ET to be correct. This checks for 287/387.
 */
check_x87:
        movb $0,X86_HARD_MATH
        clts
        fninit
        fstsw %ax
        cmpb $0,%al
        je 1f
        movl %cr0,%eax          /* no coprocessor: have to set bits */
        xorl $4,%eax            /* set EM */
        movl %eax,%cr0
        ret
        ALIGN
1:      movb $1,X86_HARD_MATH
        .byte 0xDB,0xE4         /* fsetpm for 287, ignored by 387 */
        ret