I also had the need to perform a 16bit/8bit divide. Using your routine and one I had written previously I came up with this one:
R1R0 is the 16 bit dividend
R2 is the 8 bit divisor

UIntDiv16:
mov b,#8

mov a,r2
clr c
subb A,r1
jc Exit ;Does the same thing as yours up to here

Loop2:
RotateL r1,r0
mov a,r1
subb a,r2
jc ItIs0
Itis1:
mov r1,a
ItIs0:
djnz b,Loop2

;The carry value of the subb instruction gets copied into
;r0 at the start of each loop

;At the end of the loop you have in r0 the complement of
;the result

mov a,r0
rlc a
;you rotate once more to account for the last subb

cpl a ;complement to get the result
Exit: ret

;The macro rotates XHXL to the left

Where:
RotateL Macro XH,XL
mov a,XL
rlc a
mov XL,a
mov a,XH
rlc a
mov XH,a
endm

Performing the same division with both routines I got a bit over 25% improvement in time using this one.
I hope it's not to unclear.