genirq: Delegate irq affinity setting to the irq thread
Commit-ID: 591d2fb02ea80472d846c0b8507007806bdd69cc Gitweb: http://git.kernel.org/tip/591d2fb02ea80472d846c0b8507007806bdd69cc Author: Thomas Gleixner <[email protected]> AuthorDate: Tue, 21 Jul 2009 11:09:39 +0200 Committer: Thomas Gleixner <[email protected]> CommitDate: Tue, 21 Jul 2009 14:35:07 +0200
genirq: Delegate irq affinity setting to the irq thread
irq_set_thread_affinity() calls set_cpus_allowed_ptr() which might sleep, but irq_set_thread_affinity() is called with desc->lock held and can be called from hard interrupt context as well. The code has another bug as it does not hold a ref on the task struct as required by set_cpus_allowed_ptr().
Just set the IRQTF_AFFINITY bit in action->thread_flags. The next time the thread runs it migrates itself. Solves all of the above problems nicely.
Add kerneldoc to irq_set_thread_affinity() while at it.
Signed-off-by Thomas Gleixner <[email protected]>
LKML-Reference: <new-submission>
--- include/linux/interrupt.h | 2 + kernel/irq/internals.h | 3 +- kernel/irq/manage.c | 50 +++++++++++++++++++++++++++++++++++++++----- kernel/irq/migration.c | 2 +- 4 files changed, 48 insertions(+), 9 deletions(-)
include/linux/interrupt.h +2 -0
@@ -64,11 +64,13 @@* IRQTF_RUNTHREAD - signals that the interrupt handler thread should run * IRQTF_DIED - handler thread died * IRQTF_WARNED - warning "IRQ_WAKE_THREAD w/o thread_fn" has been printed +* IRQTF_AFFINITY - irq thread is requested to adjust affinity */ enum {IRQTF_RUNTHREAD ,IRQTF_DIED ,IRQTF_WARNED , +IRQTF_AFFINITY , };typedef irqreturn_t (*irq_handler_t )(int ,void *);
kernel/irq/internals.h +1 -2
@@ -42,8 +42,7 @@ static inline void unregister_handler_proc(unsigned int irq,extern int irq_select_affinity_usr (unsigned int irq); -extern void -irq_set_thread_affinity (struct irq_desc *desc,const struct cpumask *cpumask); +extern void irq_set_thread_affinity (struct irq_desc *desc);/* * Debugging printout:
kernel/irq/manage.c +44 -6
91 lines not on this page · raw message
kernel/irq/migration.c +1 -1
@@ -45,7 +45,7 @@ void move_masked_irq(int irq) < nr_cpu_ids))if (!desc->chip->set_affinity (irq, desc->pending_mask)) {cpumask_copy (desc->affinity, desc->pending_mask); -irq_set_thread_affinity (desc, desc->pending_mask); +irq_set_thread_affinity (desc); }cpumask_clear (desc->pending_mask);