[En-Nut-Discussion] Minimalistic stack settings

Harald Kipp harald.kipp at egnite.de
Wed Jan 11 11:24:27 CET 2006


Henrik,

This information is indeed missing and I'm also using the
unreliable trial and error method with finally adding some
extra space. This is most unsatisfying.

Michael Fischer sent some code a while ago, where he initialized
the stack with specific values during thread creation and later
checked the real free space with a modified NutDumpThreadList().
Unfortunately this added too much execution time with interrupts
blocked and after some discussions we decided not to add that
code. But for testing purposes, this may be the most simple way.

Currently I do not have any idea, if the dedicated interrupt
stack introduced by Oliver Schulz is still working. At least
this will help to get a deterministic behavior of stack usage.

Harald

At 16:13 11.01.2006 +1000, you wrote:
>Hello forum,
>
>I am trying to find the smallest possible stack settings for Nut/OS 4.0.x 
>or 3.9.x which are still known to work.
>
>Currently there are 6 stack size configuration values:
>
>NUT_THREAD_IDLESTACK
>NUT_THREAD_MAINSTACK
>NUT_THREAD_TCPSMSTACK
>NUT_THREAD_SNTPSTACK
>NUT_THREAD_NICRXSTACK
>NUT_THREAD_AHDLCRXSTACK
>
>Maybe somebody tried to play with these settings and can tell me which 
>values are known to work and which settings certainly will lead to 
>unstable systems.
>
>The stack size issue matters for systems based on the AT90CAN128 chip, 
>where total stack size is limited to 4 Kbyte.
>
>Regards
>
>Henrik Maier




More information about the En-Nut-Discussion mailing list