remquo asm: return quotient mod 8, as intended by the spec

this is a lot more efficient and also what is generally wanted.
perhaps the bit shuffling could be more efficient...

src/math/i386/remquo.s

@@ -2,40 +2,49 @@
 .type remquof,@function
 remquof:
 	mov 12(%esp),%ecx
-	fldl 4(%esp)
 	fldl 8(%esp)
+	fldl 4(%esp)
+	mov 11(%esp),%dh
+	xor 7(%esp),%dh
 	jmp 1f
 
 .global remquol
 .type remquol,@function
 remquol:
 	mov 28(%esp),%ecx
-	fldl 4(%esp)
 	fldl 16(%esp)
+	fldl 4(%esp)
+	mov 25(%esp),%dh
+	xor 13(%esp),%dh
 	jmp 1f
 
 .global remquo
 .type remquo,@function
 remquo:
 	mov 20(%esp),%ecx
-	fldl 4(%esp)
 	fldl 12(%esp)
-1:	fld %st(1)
+	fldl 4(%esp)
+	mov 19(%esp),%dh
+	xor 11(%esp),%dh
 1:      fprem1
 	fnstsw %ax
 	sahf
 	jp 1b
-	fsubr %st(0),%st(2)
+	fstp %st(1)
-	fxch %st(2)
+	mov %ah,%dl
-	fdivp
+	shr %dl
-	mov $0x4f000000,%eax
+	and $1,%dl
-	mov %eax,4(%esp)
+	mov %ah,%al
-	flds 4(%esp)
+	shr $5,%al
-	fxch %st(1)
+	and $2,%al
-1:	fprem
+	or %al,%dl
-	fnstsw %ax
+	mov %ah,%al
-	sahf
+	shl $2,%al
-	jp 1b
+	and $4,%al
-	fistpl (%ecx)
+	or %al,%dl
-	fstp %st(0)
+	test %dh,%dh
+	jns 1f
+	neg %dl
+1:	movsbl %dl,%edx
+	mov %edx,(%ecx)
 	ret