[En-Nut-Discussion] NutHeapAlloc Error?

[Tim DeBaillie]

 From what I can tell, it would/could occur on ALL of the supported  
processors. The part that allows this to happen in the first place is  
that the heap size can be initialized to a NON-word boundary length.

This bug is actually quite nasty because it is VERY hard to detect. It  
is very easily confused for running out of memory, a memory leak, or  
blowing a stack in a thread. From our testing, it will occur in nearly  
ALL NutOS applications that fragment the heap or utilize nearly all of  
the heap at any point in time.

In a VERY busy system, the bug was exercised between 5-15 minutes. In  
a normal system, I would expect an application could run for weeks or  
months before exercising this case. This makes pin-pointing your  
reboot problem rather difficult!

Tim

--


On Oct 15, 2008, at 6:47 PM, "Henrik Maier" <hmlists at focus-sw.com>  
wrote:

> Just to clarify this: Can you confirm that this issue ONLY affects ARM
> architectures?
>
> Henrik
>
>> -----Original Message-----
>> Subject: Re: [En-Nut-Discussion] NutHeapAlloc Error?
>> Timothy M. De Baillie wrote:
>>> We found the error. Solution coming out soon..
>>
>> Found in Stable OS Versions 4.4.0 -> 4.6.4
>> On the ARM processors
>
> _______________________________________________
> http://lists.egnite.de/mailman/listinfo/en-nut-discussion
>