Patches for classic affinity implementation and test

Gedare Bloom gedare at rtems.org
Fri Feb 21 14:26:55 UTC 2014


On Fri, Feb 21, 2014 at 8:16 AM, Joel Sherrill
<Joel.Sherrill at oarcorp.com> wrote:
> Let me make one thing clear so we can quit discussing it. the "have sys
> cpuset" macro was a nice thing from us to avoid forcing everyone to update
> their tools immediately.
>
> There is no assumption that cpuset is not present long term and we will NOT
> provide an alternative internal implementation.
>
> In retrospect, I am beginning to think the effort being nice was wasted
> because it seems to have just drawn flak. Apparently it is more acceptable
> to just break things now.
>
Easy there Joel... get some coffee. :) The feature macro test is fine
for short-term, but we should be able to remove it for the long-term.

> --joel
>
> On Feb 21, 2014 1:58 AM, Sebastian Huber
> <sebastian.huber at embedded-brains.de> wrote:
> On 2014-02-20 21:07, Jennifer Averett wrote:
>> +#if __RTEMS_HAVE_SYS_CPUSET_H__ && defined( RTEMS_SMP )
>> +const Cpuset_Control *_Cpuset_Handler_default(void);
>> +#else
>> +#define _Cpuset_Handler_default()   do { } while ( 0 )
>> +#endif
>
> In case the C library doesn't provide an appropriate <sys/cpuset.h> then we
> should provide our own header file.  Maybe we should use a special header
> file
> that either uses <sys/cpuset.h> directly or alternatively provides the
> missing
> declarations and functions.
>
> I don't think it makes sense to use a SMP configuration with this API.
>
> --
> Sebastian Huber, embedded brains GmbH
>
> Address : Dornierstr. 4, D-82178 Puchheim, Germany
> Phone   : +49 89 189 47 41-16
> Fax     : +49 89 189 47 41-09
> E-Mail  : sebastian.huber at embedded-brains.de
> PGP     : Public key available on request.
>
> Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.
> _______________________________________________
> rtems-devel mailing list
> rtems-devel at rtems.org
> http://www.rtems.org/mailman/listinfo/rtems-devel
>
> _______________________________________________
> rtems-devel mailing list
> rtems-devel at rtems.org
> http://www.rtems.org/mailman/listinfo/rtems-devel
>




More information about the devel mailing list