On 11/06/2018 06:40 AM, Peter Zijlstra wrote:
> On Mon, Oct 15, 2018 at 04:29:30PM -0400, Waiman Long wrote:
>> +spin_lock_irq(&callback_lock);
>> +clear_bit(CS_CPU_EXCLUSIVE, &cp->flags);
>> +spin_unlock_irq(&callb
On 11/06/2018 06:37 AM, Peter Zijlstra wrote:
> On Mon, Oct 15, 2018 at 04:29:30PM -0400, Waiman Long wrote:
>> Handling error returned by update_parent_subparts_cpumask() in
>> update_cpumasks_hier() is problematic as the states may become
>> inconsistent. To avoid that and increase flexibility in
On Mon, Oct 15, 2018 at 04:29:30PM -0400, Waiman Long wrote:
> + case PRS_ENABLED:
> + if (update_parent_subparts_cpumask(cp,
> + partcmd_update, NULL, tmp))
> + update_tasks_cpum
On Mon, Oct 15, 2018 at 04:29:30PM -0400, Waiman Long wrote:
> + spin_lock_irq(&callback_lock);
> + clear_bit(CS_CPU_EXCLUSIVE, &cp->flags);
> + spin_unlock_irq(&callback_lock);
A single atomic bitop wrapped in a s
On Mon, Oct 15, 2018 at 04:29:30PM -0400, Waiman Long wrote:
> Handling error returned by update_parent_subparts_cpumask() in
> update_cpumasks_hier() is problematic as the states may become
> inconsistent. To avoid that and increase flexibility in handling other
> error cases, a new error state (-
Handling error returned by update_parent_subparts_cpumask() in
update_cpumasks_hier() is problematic as the states may become
inconsistent. To avoid that and increase flexibility in handling other
error cases, a new error state (-1) is added to the partition_root_state
flag. This new error state is
