RFC: Minimizing stack sizes on tests

Nick Withers nick.withers at anu.edu.au
Sun Jan 17 07:00:56 UTC 2016


On Fri, 2016-01-15 at 17:00 -0600, Joel Sherrill wrote:
> Hi
> 
> In the dark ages of RTEMS, the stack size argument wasn't really give
> much
> thought when writing the tests. As we move down in target size, I am
> wondering if we should make as many of the tests as possible use the
> default stack size.

Sounds like a good idea to this humble user :-P

It might even catch some that fail.

> Isaac's reduction of stack size made me think of this. It may also
> make
> sense for a BSP to define a default stack size similar to how we
> allow a
> bsp default idle stack size.
> 
> Just thinking and wanting feedback.

Just thinking here too, but I'm a bit of a fan of GCC's "-Wstack
-usage=<bytes>" feature, which'll issue a warning if it can't guarantee
that a function's stack usage is bounded and at or below <bytes> B.

I'm not really sure if it always recurses though, to be honest. The
doco (https://gcc.gnu.org/onlinedocs/gcc-4.9.3/gcc/Warning-Options.html
)'s a bit sparse...

Oh, and it isn't too great when unoptimised on constructs like 'char
buffer[strlen("bam")]' where it doesn't precompute strlen().
-- 
Nick Withers

Embedded Systems Programmer
Department of Nuclear Physics, Research School of Physics and Engineering
The Australian National University (CRICOS: 00120C)



More information about the devel mailing list