Hi, I need some info in regard to 64-bit process' stack layout.
1) Below is a daigram of how I think the stack looks like. Can anyone
tell me if this is correct, and if not, what's missing, etc.
2) As I understand stack bias serves two purposes:
one - to allow trap handlers to distinguis between 64-bit and 32-bit
register window.
two - to allow access a larger stack area with 13-bit signed constants.
Can someone show me an example how stack bias in a 64-bit process
allows to access larger stack area vs. 32-bit process with no stack
bias? Also, on the diagram below, there is an area of 2046 bytes between
%sp and %sp+BIAS, what is that area used for? Is it reserved for
something? Does is serve a purpose?
## High Address ##
---------------------------------------------------------
%fp+BIAS <-- start of previous stack frame
---------------------------------------------------------
%fp+BIAS-1 1st local variable
---------------------------------------------------------
%fp+BIAS-N 2nd local variable, etc.
---------------------------------------------------------
~~~ variable size
---------------------------------------------------------
%sp+BIAS+186 if needed, outgoing arguments 6 to ...
---------------------------------------------------------
%sp+BIAS+138 space for arguments 0 to 5 (6*8=48 bytes)
---------------------------------------------------------
%sp+BIAS+128 struct/union return pointer (8 bytes)
---------------------------------------------------------
window save area for 16 registers
%sp+BIAS (16*8=128 bytes)
---------------------------------------------------------
|
2047-1 bytes
%sp |
---------------------------------------------------------
## Low Address ##
_______________________________________________
opensolaris-code mailing list
opensolaris-code@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
