From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <456C983E.70408@domain.hid>
Date: Tue, 28 Nov 2006 21:12:46 +0100
From: Jan Kiszka <jan.kiszka@domain.hid>
MIME-Version: 1.0
Subject: Re: [Xenomai-core] Re: [RFC][PATCH 2/2] global CPU affinity control
References: <456C3EFB.7030303@domain.hid> <456C887D.9040008@domain.hid>
In-Reply-To: <456C887D.9040008@domain.hid>
Content-Type: text/plain; charset=ISO-8859-15
Content-Transfer-Encoding: 7bit
Sender: jan.kiszka@domain.hid
List-Id: "Xenomai life and development \(bug reports, patches,
	discussions\)" <xenomai.xenomai.org>
List-Unsubscribe: <https://mail.gna.org/listinfo/xenomai-core>,
	<mailto:xenomai-core-request@domain.hid>
List-Archive: </public/xenomai-core>
List-Post: <mailto:xenomai@xenomai.org>
List-Help: <mailto:xenomai-core-request@domain.hid>
List-Subscribe: <https://mail.gna.org/listinfo/xenomai-core>,
	<mailto:xenomai-core-request@domain.hid>
To: Gilles Chanteperdrix <gilles.chanteperdrix@xenomai.org>
Cc: "M. Koehrer" <mathias_koehrer@domain.hid>, xenomai-core <xenomai@xenomai.org>

Gilles Chanteperdrix wrote:
> Jan Kiszka wrote:
>> This is version 4 of the /proc/xenomai/affinity feature. It has been
>> rebased on the reworked xnshadow_map code, now applying the default
>> affinity mask in a (hopefully) correct way also on shadow threads.
>>
>> What has been said for the previous patch also applies here: test and
>> comment! Mathias, you asked for this feature, please let us know if it
>> fulfils your requirements.
>>
>> Philippe, you were looking for documentation of this new behaviour.
>> Please let me know where you would like to see this. Surely in the
>> nucleus API, I guess. Anywhere else?
>>
>> Jan
>>
>>
>> ------------------------------------------------------------------------
>> (...)
>> @@ -909,8 +911,8 @@ int xnpod_start_thread(xnthread_t *threa
>>  	if (!testbits(thread->status, XNDORMANT))
>>  		return -EBUSY;
>>  
>> -	if (xnarch_cpus_empty(affinity))
>> -		affinity = XNARCH_CPU_MASK_ALL;
>> +	if (xnarch_cpus_equal(affinity, XNPOD_ALL_CPUS))
>> +		affinity = nkaffinity;
> 
> Should not this be:
> 	if (xnarch_cpus_empty(affinity))
> 		affinity = XNARCH_CPU_MASK_ALL;
> 	xnarch_cpus_and(affinity, affinity, nkaffinity);
> 	if (xnarch_cpus_empty(affinity))
> 		return -EINVAL;
> 
> so that no thread could escape the rule imposed by nkaffinity.
> 

Nope, that's intentional (forgot to repeat this fact, was explained
earlier): Users can always override nkaffinity by their own mask. Only
the "I-don't-care" XNPOD_ALL_CPUS activates this new selection mechanism.

Jan