Hi,
I have been writing codes in assembly for a long time now. From 8051 to AVR now.
But i wish to make my codes more organized, more understandable.
I do use comments and use EQU mappings but in good organized codes, i see coders include device drivers, macros and definations like .eseg, .dseg, .cseg.
is there any tutorial available on the proper usage of .eseg, .dseg, .cseg :roll:
any help and link to tutorials will be greatly appreciated.
thanks :)
js 
any tutorial available on the proper usage of .eseg, .dseg, .cseg
.cseg is the flash area where your code resides
.dseg is your ram where your variables reside
.eseg is the EEPROM area where data can be stored.
JohanEkdahl 
is there any tutorial available on the proper usage of .eseg, .dseg, .cseg
When everything else fails, read the documentation.. :wink:
http://www.atmel.no/webdoc/avras...
If you have Atmel Studio installed you should have similar docs on your hard disk.
clawson 
But i wish to make my codes more organized, more understandable.
(someone had to say it! ;-))
hi js
How did you get your code to work if you don't know the meaning of these?
I have been writing codes like this,
.include "m16def.inc" .equ FCPU = 7372800 .equ DRly = 0 .equ C1RlyU = 1 .equ C1RlyD = 2 ldi r16,Low(RamEnd) out SPL,r16 ldi r16,High(RamEnd) out SPH,r16 ldi temp,((FCPU/0x2580)/16)-1 out UBRR,temp ldi temp,0b00011000 out UCSRB,temp ldi temp,0b11111111 out DDRC,temp out DDRD,temp . . . . . . CMD1: .db "[D1O]",0 CMD2: .db "[D1C]",0 CMD3: .db "[C1O]",0 CMD4: .db "[C1C]",0
JohanEkdahl, thanks for the document. but i wanted to see how it is used in a code
Switch to C?
clawson, No i am more comfortable with ASM and that too im doing it for almost 10 years now :)
js I have been writing codes like this,
;==================== .DSEG ;==================== EEP_ADDR: .byte 2 ;Pointer to data in eeprom internal or external scroll_time: .byte 1 ;Gets copied to scroll_delay every scan scroll_delay: .byte 1 ;Number of scan delay msg0: .byte PKIO_DATA_SIZE+1 ;Message 0 in ram.1 byte larger for CR msg0_index: .byte 2 ;Pointer to next char in msg 0 buffer dwell_timer: .byte 1 ;Beginning or End of messgae dwell timer op2_timer: .byte 1 ;Timer value for OP2 op2_timer_buf: .byte 1 ;Buffer for timer value for OP2 op1_timer: .byte 1 ;Timer value for OP1 op1_timer_buf: .byte 1 ;Buffer for timer value for OP1 DIS_BUF: .byte 26 ;Display buffer+1 for scroll char msg_bfr: .byte 26 ;Current 25 + scroll chars being displayed spi_out_bfr: .byte 21 ;10 lots of 2 bytes (+1) sorted to send out cursor: .byte 2 ;Cursor position pkt_type: .byte 1 ;Type of communication packet
or put data in EEPROM
;==================== .ESEG ;==================== ;EEPROM Area nothing: ;Don't use first location .db 0 display_addr: .db 0x1a ;UART's default baud rate. MUST STAY IN THIS ORDER. _ubrrl: .db low(fosc/(16*baud)-1) _ubrrh: .db high(fosc/(16*baud)-1) ;Type of communication packet ;0=CR control type ;1=RS485 type _pkt_type: .db 0x00 ;Dwell time at beginning of message in 100mS steps beg_dwell: .db 20 ;2 Seconds ;Dwell time at end of message in 100mS steps end_dwell: .db 15 ;1.5 Seconds
david.prentice 
@saumya73,
Most importantly, observe how Mr Samperi uses explanatory comments.
IMHO, textbooks are not that good at promoting ASM style. You learn most be reading other people's mistakes / success. Avoid the bad bits and steal the good ideas.
Oh, and my first step is to format the code so that it is pleasant to read.
David.
Hi JS,
Thanks for the code example, here is how i use RAM and EPROM
.def sbuf = r16 .equ STX = 0x5B .equ ETX = 0x5D WaitReq: ldi YL,0xB0 ;Point YL Location ldi YH,0x00 ;Point YH Location WaitReq1: call RcvChar ; Get data from UART cpi Sbuf,STX ; Compare with Start Byte brne WaitReq1 ; If no start byte wait again st Y+,sbuf ;Start byte is received, store it and increase Y location GetSer: call rcvchar ;Receive next byte and st Y+,sbuf ;save it to location pointed by Y cpi sbuf,ETX ;untill End Charecter is received brne GetSer
Another way I write
;======================= ; RAM Read / Write ;======================= ;============================ ; READ from RAM ADDRESS 0x2F ;============================ lds r16,0x2F ; Load value in r16 from RAM location in 0x2F ;OR .equ Date = 0x2F lds r16,Date ; Load value in r16 from RAM location ;============================ ; WRITE to RAM ADDRESS 0x2F ;============================ sts 0x2F,r16 ; Save it to Ram Location 0x2F ;OR .equ Date = 0x2F sts Date,r16 ; Save it to Ram Location 0x2F (Date)
for EEPROM i write like this
;============================ ; WRITE to EEPROM ADDRESS 0x12 ;============================ .def edata = r16 .def addr = r17 ldi addr,0x12 ;Point Address to location 0x12 in EEPROM ldi edata,0x18 ;Load Data with constant call e2pw ; Write to EEPROM ;OR .def edata = r16 .def addr = r17 .equ Time = 0x12 ldi addr,Time ;Point Address to location 0x12 in EEPROM ldi edata,0x18 ;Load Data with constant call e2pw ; Write to EEPROM
;====================
.DSEG
;====================EEP_ADDR: .byte 2 ;Pointer to data in eeprom internal or external
scroll_time: .byte 1 ;Gets copied to scroll_delay every scan
Okay i think i can understand this, you are reserving two bytes for EEP_ADDR one byte for scroll_time
now what is the ram location for EEP_ADDR and scroll_time, and how to i access data to/from them?
is it that the assembler knows the location when we select the device? (Oh i forgot to mention i am using AVR Studio 4.18 )
for .ESEG part, I am not sure, I dont understand anything. Please let me understand it myself before i post a silly question.
@ david
You are absolutely right, commenting is very helpful in code debugging and easy understanding of code flow, but to be very frank, I alone have to manage a lot of things together from designing PCBs, writing codes, assembling PCBs, product marketing, understanding end user requirements, followup for payments and the list goes on and on.. :)
now the funy thing i get lazy while making comments on my code, then i think I know what i am writing! why comment? But yes again, I do admit that commenting is necessary and will start doing that very soon. thanks for the guide
js what is the ram location for EEP_ADDR and scroll_time,
i am using AVR Studio 4.18
JohanEkdahl
and how to i access data to/from them?
is it that the assembler knows the location when we select the device?
So, assuming you have no previous .DESG sections in your source, EEP_ADDR will be allocated starting at byte 0, and scroll_time starting at byte 2. The assembler will keep a table of symbolic names and coresponding addresses, and when you have coded e.g. a load/read instruction using the symbolic name scroll_time the assembler will know that it should actually use the address 2 in that instruction.
clawson Just think of .cseg, .dseg and .eseg as maintaining separate .org counters. By default if you don't actually specify one you get .cseg (code segment) so you can position stuff with .org then each opcode you write automatically increments the .xseg counter by the length of that opcode. .dseg (data segment) is for data, .eseg (eeprom segment) is for eeprom. Each have their own .org counters. You can position to a specific location with something like:
.eseg .org 123 data: .db 55
and that will put 55 in location 123 of the EEPROM (in the .eep file which you then upload to the chip to prepopulate the EEPROM contents). If you:
.eseg .org 123 data: .db 55 bigger: .dw 12345
Then the 55 is still written to location 123 but the 12345 value is then automatically written to 124 and 125 (0x39 in 124 and 0x30 in 125) - the .eseg org counter has automatically stepped on as it knows a .db takes one byte. After the .dw the location counter will be 126 which is where anything else you define will be located. You can also do this:
.cseg .org 0 rjmp reset .eseg .org 123 data: .db 55 bigger: .dw 12345 .cseg .org INT0_addr rjmp INT0_handler .org 10 reset: ldi r16, low(RAMEND) ...
You start in .cseg (actually I didn't need to type that as it's the default anyway) but after putting an RJMP at 0 I then happened to decide (an odd choice perhaps!) to switch to the EEPROM segment and define some data there). After that I switch back to the .cseg and the location counter is where I last left it - just after that first RJMP but I decide to .org to the address of the INT0 vector and put a jump to its handler there. I then arbitrarily chose to .org to location 10 (I'm hoping that's beyond the INT0 vector!) and continue the reset code from there on.
When it comes time to access the data in EEPROM I don't actually need to worry about 123/124/125 as the value (55 and 12345) have been given names. If I want to do a single byte EEPROM read to get the 55 value I just load the address register with "data" and the assembler will know that is 123 so will load 123 into the register.
If I want to make a double byte read I can read one byte from "bigger" and one byte from "bigger + 1" and put them in adjacent registers (like (bigger)->R26 and (bigger + 1)->r27 then the combined r27:r26 hold 12345 and if I "ADIW R26,1" then the pair will contain 12346 and so on.
Hey!!! Thanks a lot everyone, this tutorial have been very instructive.
I have written a small dummy code using .cseg, .dseg and .eseg directive.
include "m16def.inc"
.def var1 = r16
.def var2 = r17
.cseg
.org 0x0000
rjmp Reset
.org 0x0001
rjmp ExInt0
Reset:
ldi r16,Low(RamEnd)
out spl,r16
ldi r16,High(RamEnd)
out sph,r16
endless:
jmp endless
ExInt0:
reti
.dseg
Date: .byte 1 ;Reserve 1 byte for Date (address 60)
Month: .byte 1 ;Reserve 1 byte for Month (address 61)
Year: .byte 1 ;Reserve 1 byte for Year (address 62)
ADC_Value .byte 2 ;Reserve 2 byte for ADC Value (address 63 and 64)
UartData .byte 20 ;Reserve 20 byte for UART Data (address 65 to 85)
Index .byte 1 ;Reserve 1 byte for Index (address 86)
.eseg
CalVal: .db 0x21, 0x5F ;Save default Calibration bytes to EEPROM Location 00 and 01
DefSet: .db 0x0A ;Save default startup settings to EEPROM Location 02
.org 0x2F
NewSet: .db 0x32 ;Save NewSet Values to EEPROM Location 0x2F
have i used the above definitions correctly?
if it is correct, to access the Date, should I write
lds r16,Date
ldi r16,0x55
sts index,r16
ldi ZL,low(2*CalVal)
ldi ZH,high(2*CalVal)
;Here ZL,ZH is pointing to EEPROM Location 00
ldi ZL,low(2*NewSet)
ldi ZH,high(2*NewSet)
;Here ZL,ZH is pointing to EEPROM Location 2F
please correct me if im wrong some where.
js EEPROM is BYTE mapped not word mapped like flash
ldi ZL,low(2*CalVal)
ldi ZH,high(2*CalVal)
;Here ZL,ZH is pointing to EEPROM Location 00
don't use 2* that's only needed for flash
edit returning from breakfast interrupt request....
look at the data sheet as it has example code on how to read and write to the EEPROM
look at the data sheet as it has example code on how to read and write to the EEPROM
hi John, I know accessing the EEPROM, just wanted a more descriptive way to access them :)
Thanks a lot!!
Hi All,
Just a discussion (out of this thread topic)
I see most people use C to program AVRs while i prefer Assembly. Even 90% of code examples and questions over the net is based on C
Is C in any case better than Assembly :?:
Just woke up with a bad dream that I will be stuck in future devices like XMega or ARM. :roll:
JohanEkdahl
Is C in any case better than Assembly
Just make up your own mind and go with that.
Or pose a VERY specific question regarding assembler v/s C. (No "Is assembler more efficient?" is not by far specific enough.)
Just woke up with a bad dream that I will be stuck in future devices like XMega or ARM. :roll:
david.prentice Hi All,Just a discussion (out of this thread topic)
I see most people use C to program AVRs while i prefer Assembly. Even 90% of code examples and questions over the net is based on C
Is C in any case better than Assembly :?:
Just woke up with a bad dream that I will be stuck in future devices like XMega or ARM. :roll:
Stick with whichever language you are most comfortable with.
Regarding efficiency, most tasks involve setting bits in SFRs. Then checking/waiting for other bits in SFRs. You are governed by the time that it takes the peripheral to complete its task. The HLL generates the same SBI or SBIC as you would use in assembly.
Of course, regular maths, loops etc can be coded more efficiently in ASM. In practice, it is only one or two hotspots that will make any difference.
Only you can assess the development time versus benefit of anything.
David.
clawson
Even 90% of code examples and questions over the net is based on C
Apart from a copy of the invaluable Kerninghan and Ritchie that anyone writing C should have I always found the C library reference for the Borland C compiler to be essential too. It not only showed all the standard functions in the C library but had a very useful example of each. The modern equivalent that I still use regularly is:
http://www.cplusplus.com/reference/
here's just a random example of one of the many standard C functions it documents:
http://www.cplusplus.com/referen...
Although the list of reference material is focussed on C++ rather than C the fact is that all the standard C libraries
No hard feelings, No hurting anyone's sentiment
No fights, No debates.. :P
I prefer Assembly.. and will stick to it as much as possible :)