[PATCH for-4.8] irq/affinity: use get/put_online_cpus

[PATCH for-4.8] irq/affinity: use get/put_online_cpus

[Christoph Hellwig]

Without locking out CPU mask operations we might end up with an inconsistent
view of the cpumask in the function.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 kernel/irq/affinity.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
index f689593..32f6cfc 100644
--- a/kernel/irq/affinity.c
+++ b/kernel/irq/affinity.c
@@ -39,6 +39,7 @@ struct cpumask *irq_create_affinity_mask(unsigned int *nr_vecs)
 		return NULL;
 	}
 
+	get_online_cpus();
 	if (max_vecs >= num_online_cpus()) {
 		cpumask_copy(affinity_mask, cpu_online_mask);
 		*nr_vecs = num_online_cpus();
@@ -56,6 +57,7 @@ struct cpumask *irq_create_affinity_mask(unsigned int *nr_vecs)
 		}
 		*nr_vecs = vecs;
 	}
+	put_online_cpus();
 
 	return affinity_mask;
 }
-- 
2.1.4

Re: [PATCH for-4.8] irq/affinity: use get/put_online_cpus

[Christoph Hellwig]

Any chance to get a quick review for this and get it off to Linus?
Without this fix the function is unusable in 4.8+ on my test systems
because we seem to always get a inconsistent view of the cpu online
bitmap - something that didn't happen before this merge window oddly
enough.

Re: [PATCH for-4.8] irq/affinity: use get/put_online_cpus

[Thomas Gleixner]

On Sun, 14 Aug 2016, Christoph Hellwig wrote:
> Any chance to get a quick review for this and get it off to Linus?
> Without this fix the function is unusable in 4.8+ on my test systems
> because we seem to always get a inconsistent view of the cpu online
> bitmap - something that didn't happen before this merge window oddly
> enough.

That only happens if you have concurrent cpu hotplug operations, but yes this
needs to be fixed anyway.

Thanks,

	tglx

[tip:irq/urgent] genirq/affinity: Use get/put_online_cpus around cpumask operations

[tip-bot for Christoph Hellwig]

Commit-ID:  3ee0ce2a54dff07d09440723594df89bc1a12e79
Gitweb:     http://git.kernel.org/tip/3ee0ce2a54dff07d09440723594df89bc1a12e79
Author:     Christoph Hellwig <hch@lst.de>
AuthorDate: Thu, 11 Aug 2016 07:06:45 -0700
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Mon, 22 Aug 2016 11:22:44 +0200

genirq/affinity: Use get/put_online_cpus around cpumask operations

Without locking out CPU mask operations we might end up with an inconsistent
view of the cpumask in the function.

Fixes: 5e385a6ef31f: "genirq: Add a helper to spread an affinity mask for MSI/MSI-X vectors"
Signed-off-by: Christoph Hellwig <hch@lst.de>
Link: http://lkml.kernel.org/r/1470924405-25728-1-git-send-email-hch@lst.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

---
 kernel/irq/affinity.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/irq/affinity.c b/kernel/irq/affinity.c
index f689593..32f6cfc 100644
--- a/kernel/irq/affinity.c
+++ b/kernel/irq/affinity.c
@@ -39,6 +39,7 @@ struct cpumask *irq_create_affinity_mask(unsigned int *nr_vecs)
 		return NULL;
 	}
 
+	get_online_cpus();
 	if (max_vecs >= num_online_cpus()) {
 		cpumask_copy(affinity_mask, cpu_online_mask);
 		*nr_vecs = num_online_cpus();
@@ -56,6 +57,7 @@ struct cpumask *irq_create_affinity_mask(unsigned int *nr_vecs)
 		}
 		*nr_vecs = vecs;
 	}
+	put_online_cpus();
 
 	return affinity_mask;
 }