Re: [avr-gcc-list] Why EthNut2.4.2 should use -Os instead of -O3?

[ken restivo]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Thanks. I'd been wondering about that.  

Adding 

        void NutThreadSwitch(void) __attribute__ ((noinline));

Seems to get everything to at least compile with -O3, using CVS avr-gcc 3.3 and 
avr-libc from a few days ago.

In the thread.lst file, it looks like thread_start: is in the right place, but 
I don't (yet) have a target to test it on.

- -ken
- ------
On Fri, Aug 30, 2002 at 10:31:18AM +0200, Harald Kipp wrote:
> The Ethernut RTOS inherits an adventurous
> construction from the original Liquorice
> source code:
> 
> static thread_switch()
> {
>   ...
>   asm("label:");  /* initial thread entry */
>   ...
> }
> 
> create_thread()
> {
>   ...
>   asm("jmp label");
>   ...
> }
> 
> When using -o3 with AVRGCC 3.2, the compiler
> makes thread_switch() inline. This makes 'label'
> ambiguous.
> 
> Harald
> 
> avr-gcc-list at http://avr1.org

- -- 
- ---------------
The world's most affordable web hosting.
http://www.nearlyfreespeech.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE9cBrCe8HF+6xeOIcRAslQAJ4tJQYoeq8pG//a7L1h+0VSKXx9+ACfWY5K
c/m9+5Y8oBJoqGMb1XZC3AA=
=4cta
-----END PGP SIGNATURE-----
avr-gcc-list at http://avr1.org