inline functions

[Gerke Kok]

This heap thing has bothered my target as well... I postponed it by setting the heap-start at a verry alligned address. But when I now hear some other people also have problems I start to thing this might be my problem as well. Still hoping for a solution...
wkr,
Gerke


*********** REPLY SEPARATOR  ***********

On 07/03/02 at 17:01 Gerald Needell wrote:

>Can somone tell me if I'm barking up the wrong tree here.
>My saga began sevearl posts back with problems  using the "debug" 
>version of RTEMS. In digging, I have found that my application is dying 
>in when trying to release heap space. The code involved uses several 
>"inline" functions (..c/src/exec/score/inline/rtems/score/heap.inl).  In 
>reading the gcc docs, I found that when optimization is turned off, then 
>so is "inlining" so all of these functions are going in as normal code. 
>My question is, is this liketly to be causing my problems? The failure 
>seems to occur when the address of a heap block is calculated and it 
>times up non-word aligned. I am trying to persue this in two directions. 
>One, the excess stack usage by these function not being inline is 
>causing the problem and I'm clobbering a pointer or two, the unoptimized 
>compilation of the functions is produing a legitimate alignment problem.
>
>  So the reason for the post is can anyone tell me if this makes any 
>sense or do you have any suggestions.
>
>thanks - Jerry
>
>
>-- 
>_______________________________________________________________________
>Jerry Needell                     | Internet: jerry.needell at unh.edu
>Space Science Center/Morse Hall   | Voice: (603) 862 2732
>University Of New Hampshire       | FAX:   (603) 862 0311
>Durham, NH 03824                  |