On Wed, 4 Jun 2008, Paul Jackson wrote:

quoted text
> Do we need a new PF_* flag for this?  Perhaps one can test for this
> by examining the currently available properties of tasks.  Would it
> be sufficient to look for kernel threads (NULL mm_struct) whose
> cpus_allowed is a strict subset of the online CPUs?
> 

That would only identify kthreads that have been created with a subsequent 
call to set_cpus_allowed() or kthread_bind().

The PF_CPU_BOUND change targets only the latter since there are kthreads, 
such as kstopmachine, that can continue to manipulate their cpus_allowed 
during their lifetime.

Other kthreads such as the scheduler migration thread and soft lockup 
watchdog thread, however, always stay bound to a single cpu by use of 
kthread_bind().  These are the tasks that get the PF_CPU_BOUND flag and 
cannot be rebound via set_cpus_allowed() because of their negative 
effects.

		David
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/