[avr-libc-dev] Re: [bug #19445] _malloc_heap_end does not follow _heap_end on m2561

[Joerg Wunsch]

As Steve Franks wrote:

> FYI, We are having serious issues (reboots, etc) on the project that
> generated this bug report, and cannot appear to find any
> pointer/bounds issues that would cause that type of behavior.

However, just to remind you, the subject of this bug report is
*solely* to fix the slightly misleading picture in the docs that made
you think __malloc_heap_end were an address when it is actually a
variable pointing to an address.

> Of course, we are building on an arbirary version of avr-libc from
> cvs (that we pulled in to fix the avr6/sprintf bug), so maybe some
> other isses are there due to that.

At least the malloc() stuff has not been changed in ages (more than
1.5 years ago).

Alas, there's not much I could see to help you debugging that.  Did
you try analyzing the stack/heap usage?  (By filling the entire area
with a pattern initially, and then looking how much of that pattern
has been overwritten.)

-- 
cheers, J"org               .-.-.   --... ...--   -.. .  DL8DTL

http://www.sax.de/~joerg/                        NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)