Reading integers > 9 in x86 assembly

I want to read integers from input in x86 assembly but have some problems doing so when the integer is bigger than 9.

I tried the following code (found in internet) :

.code
  mov bh,0
  mov bl,10
inputloop:
  mov ah,1
  int 21h
  cmp al,13
jne convertion
jmp startcalc

convertion:
  sub al,48
  mov cl,al
  mov al,bh
  mul bl
  add al,cl
  mov bh,al
jmp inputloop    

startcalc:

I want my program to store the correct number in ax register at the start of startcalc label.

What should I do and what should I change in this program?

Oh! My! That code limits your input to a number within 1 byte, so between 0 and 255.

Using bx to store the result, I would use something like this:

  mov bx, 0

inputloop:

  mov ah, 1   ; getc()
  int 21h

  cmp al, 13  ; enter?
  je startcalc

  sub al, 48  ; character to number (assuming user only types '0' to '9')
  mov ah, 0   ; clear ah

  shl bx, 1   ; x2
  mov cx, bx
  shl bx, 2   ; x4 (so x8 total)
  add bx, cx  ; x2 + x8 = x10
  add bx, ax

  jmp inputloop

startcalc:
   mov ax, bx  ; if you want the result in ax instead

Now your number is between 0 and 65535, at least.

Note: the shl and add to do the "bx × 10" could be replaced by an imul on newer processors since those run that instruction in three cycle. Much older processors were really slow and the 4 instructions above were more than twice faster (or 5 instructions on 8086 where shl did not accept an immediate other than 1). As Peter mentioned below, the lea trick can also be used.