Reading specific byte in memory without wasting code space - 8051

Question

I'm trying to come up with a more efficient method to read a series of bytes from extended memory in the AT89S52 microcontroller.

I have organized memory storage in such as way where the ram position is dependant on "bank" and "Unit". I'll illustrate with defines:

Storage1 equ 0C000h ;bank 1 for unit 1 = C0:00h to C0:FFh
                    ;bank 1 for unit 2 = C1:00h to C1:FFh
                    ;bank 1 for unit 3 = C2:00h to C2:FFh
                    ...
                    ;bank 1 for unit 16 = CF:00h to CF:FFh

Storage2 equ 09000h ;bank 2 for unit 1 = 90:00h to 90:FFh
                    ;bank 2 for unit 2 = 91:00h to 91:FFh
                    ;bank 2 for unit 3 = 92:00h to 92:FFh
                    ...
                    ;bank 2 for unit 16 = 9F:00h to 9F:FFh

Here, each unit for each bank has 256 bytes. Currently I'm able to access each unit with the following code:

StorageN equ 0C000h
UnitNo equ 2h

mov DPTR,#StorageN
mov A,#UnitNo
orl DPH,A
movx A,@DPTR

After assessing this code with online documentation from Keil, this code requires 8 bytes of storage space.

For readability, I also have defines where special bytes of each unit in each storage bank have special meaning as so:

Storage1Status equ 0C0FEh ;255th byte of all units in bank1=status
Storage2Status equ 0C0FBh ;252nd byte of all units in bank2=status

If I were to take the status of storage unit 2 from bank 1 and put the result into storage unit 5 bank 2, I would use the following code:

mov DPTR,#Storage1Status
mov A,#1h ;unit 1
orl DPH,A
movx A,@DPTR
mov R7,A ;save status

mov DPTR,#Storage2Status
mov A,#5h ;unit 5
orl DPH,A
mov A,R7
movx @DPTR,A

That takes about 18 bytes of code space each time to update one value from one location to another.

Much of the time in my application, the unit numbers stay the same but the banks change often. so I'd use code like this to access the banks:

mov DPTR,#Storage2
mov A,#1h ;unit 1
orl DPH,A
movx A,@DPTR

mov DPTR,#Storage1
mov A,#1h ;unit 1
orl DPH,A
movx A,@DPTR

mov DPTR,#StorageN
mov A,#1h ;unit 1
orl DPH,A
movx A,@DPTR

...

mov DPTR,#StorageX
mov A,#1h ;unit 1
orl DPH,A
movx A,@DPTR

But thats a lot of code bytes being wasted just to take one byte from different storage banks. Is there a way to perform such processing without needing so much storage space?