]> git.ipfire.org Git - thirdparty/linux.git/commitdiff
block: Protect against concurrent isolated cpuset change
authorFrederic Weisbecker <frederic@kernel.org>
Thu, 15 May 2025 14:50:04 +0000 (16:50 +0200)
committerFrederic Weisbecker <frederic@kernel.org>
Tue, 3 Feb 2026 14:23:33 +0000 (15:23 +0100)
The block subsystem prevents running the workqueue to isolated CPUs,
including those defined by cpuset isolated partitions. Since
HK_TYPE_DOMAIN will soon contain both and be subject to runtime
modifications, synchronize against housekeeping using the relevant lock.

For full support of cpuset changes, the block subsystem may need to
propagate changes to isolated cpumask through the workqueue in the
future.

Signed-off-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Jens Axboe <axboe@kernel.dk>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Marco Crivellari <marco.crivellari@suse.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tejun Heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Waiman Long <longman@redhat.com>
Cc: linux-block@vger.kernel.org
block/blk-mq.c

index 1978eef95dca3fb332a73aeff7b9613ee770a8a3..0037af1216f35771c722610b78c279ed67359bfe 100644 (file)
@@ -4257,12 +4257,16 @@ static void blk_mq_map_swqueue(struct request_queue *q)
 
                /*
                 * Rule out isolated CPUs from hctx->cpumask to avoid
-                * running block kworker on isolated CPUs
+                * running block kworker on isolated CPUs.
+                * FIXME: cpuset should propagate further changes to isolated CPUs
+                * here.
                 */
+               rcu_read_lock();
                for_each_cpu(cpu, hctx->cpumask) {
                        if (cpu_is_isolated(cpu))
                                cpumask_clear_cpu(cpu, hctx->cpumask);
                }
+               rcu_read_unlock();
 
                /*
                 * Initialize batch roundrobin counts