Author Topic: what is frame pointer and pushl???  (Read 4818 times)

nobody

  • Guest
what is frame pointer and pushl???
« on: January 21, 2005, 11:13:18 AM »
hi there i was just asking what is the function of frame pointer especially in stack can u give me links, although i ve googled but didnt find a good explanation.
Second what is pushl??? cause i ve studied an assembly course but didnt find this "pushl" all i know is push, pushA, pushAD, pushf...etc
Thanks alot.
0X66FE

Offline Frank Kotler

  • NASM Developer
  • Hero Member
  • *****
  • Posts: 2437
  • Country: us
Re: what is frame pointer and pushl???
« Reply #1 on: January 21, 2005, 12:21:04 PM »
The "frame pointer" gives you access to parameters passed on the stack, and to local variables, usually using small (signed byte) offsets (smaller opcodes). Best place to learn more is probably http://www.drpaulcarter.com/~pcasm

"Pushl" looks like at&t (Gas) syntax, not Nasm. It just means a 32-bit push - "push dword ..." in Nasm syntax. FWIW, in Nasm 0.98, "push 1" was not valid. You *had* to specify either "push word 1" or "push dword 1". Shortly after the release of 0.98, the folks involved in Nasm development at that time - primarily H. Peter Anvin (still with us) and John Fine (gone on to other interests, and sorely missed) - decided that it would be sensible for Nasm to "assume" that the size of a "push" was word if Nasm was emitting in "bits 16" mode, and dword if in "bits 32". All later versions of Nasm work this way - you can override it if you need to, but this is extremely rare!

To further complicate the issue, there's "push byte 1". This does *not* push a byte on the stack - there's no such instruction. The operand is *stored* as a byte in the code stream, but expanded - sign-extended - to word or dword when it's pushed. Nasm does not make this optimization by default - you have to specify "push byte" (the apparently contradictory "push dword byte" works, too - Nasm knows what you mean). Using the "-O" switch (use a good big "maximum passes" parameter - I like "-O999" - add a "v" for "verbose" if you want a report on how many passes it actually used) will cause Nasm to use the "push byte" form if the supplied operand will fit in a signed byte - -128 thru +127.

Best,
Frank

nobody

  • Guest
Re: what is frame pointer and pushl???
« Reply #2 on: January 21, 2005, 12:42:26 PM »
Thanks man for ur help :)
0X66FE