Linux block layer
 help / color / mirror / Atom feed
* [PATCH v5] blk-mq: dispatch kworker to housekeeping CPU when hctx cpumask is empty
@ 2026-11-21 23:52 Aaron Tomlin
  0 siblings, 0 replies; only message in thread
From: Aaron Tomlin @ 2026-11-21 23:52 UTC (permalink / raw)
  To: axboe
  Cc: tom.leiming, hare, dwagner, emilne, frederic, marco.crivellari,
	tj, atomlin, sean, linux-block, linux-kernel, stable

Commit a46c27026da1 ("blk-mq: don't schedule block kworker on isolated
CPUs") excluded isolated CPUs from hctx->cpumask to prevent worker
threads from executing block work on isolated CPUs. When all CPUs mapped
to a hardware queue are isolated (i.e. blk_mq_hctx_empty_cpumask(hctx)
evaluates to true), blk_mq_hctx_next_cpu() returns WORK_CPU_UNBOUND with
the intention of offloading execution.

However, kblockd_workqueue is allocated as a per-CPU workqueue
(WQ_PERCPU). When WORK_CPU_UNBOUND is queued on a per-CPU workqueue, the
workqueue core resolves the target CPU via raw_smp_processor_id().
Consequently, if a task executing on an isolated CPU submits I/O and
triggers an asynchronous queue dispatch with zero delay (for instance,
upon hardware queue congestion, tag/resource exhaustion returning
BLK_STS_RESOURCE, or plugging), &hctx->run_work is queued directly onto
the local isolated CPU's worker pool. This schedules the local per-CPU
high-priority kworker (such as kworker/6:1H-kblockd) on the isolated
CPU. In real-time environments, an active SCHED_FIFO workload will
starve this SCHED_NORMAL worker and stall hardware dispatch, whilst also
introducing unwanted latency jitter on nohz_full cores, defeating that
commit's original objective.

Introduce blk_mq_hctx_unbound_cpu() to select an online housekeeping CPU
from housekeeping_cpumask(HK_TYPE_DOMAIN) via
cpumask_any_and_distribute() when executing on an isolated CPU, and
otherwise return WORK_CPU_UNBOUND. Gate the helper behind the static key
housekeeping_overridden under CONFIG_CPU_ISOLATION so that common
configurations without CPU isolation incur zero overhead across
single-queue devices, hardware queues with empty cpumasks, and offline
CPU fallback paths, whilst non-isolated CPUs preserve local execution
via WORK_CPU_UNBOUND. Guard the slow-path check and cpumask selection
with guard(preempt)() to pin raw_smp_processor_id(), satisfy RCU
dereferencing of housekeeping_cpumask(HK_TYPE_DOMAIN), and protect the
per-CPU distribution state.

Fixes: a46c27026da1 ("blk-mq: don't schedule block kworker on isolated CPUs")
Cc: Ming Lei <tom.leiming@gmail.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Frederic Weisbecker <frederic@kernel.org>
Cc: Marco Crivellari <marco.crivellari@suse.com>
Cc: stable@vger.kernel.org
Signed-off-by: Aaron Tomlin <atomlin@atomlin.com>
---
Note: As Frederic Weisbecker pointed out on v4, workqueue-level support
(WQ_PREFER_PERCPU) is currently in development to handle isolated CPUs
automatically, at which point kblockd_workqueue will be a candidate for
simplification; in the meantime, this provides a minimal fix suitable
for backporting.

Changes since v4:

 - Rebased against v7.4-rc1

 - Hoisted #ifdef CONFIG_CPU_ISOLATION and
   static_branch_unlikely(&housekeeping_overridden) into
   blk_mq_hctx_unbound_cpu() so blk_mq_hctx_next_cpu() retains a single
   fast-path check and non-isolated configurations avoid any RCU or
   preemption guards across all unbound fallback paths

 - Moved guard(preempt)() ahead of cpu_is_isolated(raw_smp_processor_id())
   to prevent migration between reading the local CPU ID and selecting a
   housekeeping CPU under CONFIG_PREEMPT_RCU, whilst satisfying
   rcu_dereference_all_check() without a separate guard(rcu)()

 - Consolidated the commit message summary into prose

 - Link to v4: https://lore.kernel.org/lkml/20261022034737.47820-1-atomlin@atomlin.com/

Changes since v3:

 - Create companion blktests test case (block/050) to reproduce this issue
   and verify the fix (Daniel Wagner)

 - Gate the single-queue (nr_hw_queues == 1) check behind the static key
   housekeeping_overridden under CONFIG_CPU_ISOLATION to ensure the
   common case without CPU isolation incurs zero overhead on the fast
   path (Jens Axboe)

 - Wrap the evaluation of cpu_is_isolated() inside guard(rcu)() in
   blk_mq_hctx_unbound_cpu() to prevent an RCU lockdep warning and ensure
   safe dereferencing of the housekeeping cpumask while concurrent cpuset
   updates occur

 - Disable preemption via guard(preempt)() around cpumask_any_and_distribute()
   to prevent a preemption warning under CONFIG_PREEMPT_RCU and avoid
   data races when accessing per-CPU distribution state

 - Link to v3: https://lore.kernel.org/all/20261016052203.489760-1-atomlin@atomlin.com/

Changes since v2:

 - Query housekeeping_cpumask(HK_TYPE_DOMAIN) instead of HK_TYPE_TIMER
   to align with cpu_is_isolated(), workqueue affinity rules, and cpuset
   isolated partitions

 - Query the housekeeping cpumask under guard(rcu)() to safely handle
   dynamic cpuset updates without lockdep warnings

 - Validate that cpumask_any_and_distribute() returns a valid cpu ID
   before returning, falling back cleanly to WORK_CPU_UNBOUND

 - Link to v2: https://lore.kernel.org/lkml/20261015020225.422107-1-atomlin@atomlin.com/

Changes since v1:

 - Introduce the blk_mq_hctx_unbound_cpu() helper to steer deferred work
   to an online housekeeping CPU only when executing on an isolated
   core, preserving WORK_CPU_UNBOUND and local cache locality for
   housekeeping cores

 - Apply blk_mq_hctx_unbound_cpu() across all unbound fallback paths in
   blk_mq_hctx_next_cpu(), comprehensively covering single-queue devices
   (nr_hw_queues == 1), empty hardware queue cpumasks, and the offline
   CPU fallback path.

 - Elaborate in the commit log on the triggers for asynchronous queue
   dispatch (e.g. tag exhaustion, BLK_STS_RESOURCE backpressure, and
   plugging) and the SCHED_FIFO worker starvation hazard on isolated
   cores

 - Link to v1: https://lore.kernel.org/lkml/20261012012905.267920-1-atomlin@atomlin.com/
---
 block/blk-mq.c | 23 +++++++++++++++++++++--
 1 file changed, 21 insertions(+), 2 deletions(-)

diff --git a/block/blk-mq.c b/block/blk-mq.c
index d9ab295be4d9..4657a3287f3a 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -2252,6 +2252,25 @@ static bool blk_mq_hctx_empty_cpumask(struct blk_mq_hw_ctx *hctx)
         return hctx->next_cpu >= nr_cpu_ids;
 }
 
+static inline int blk_mq_hctx_unbound_cpu(void)
+{
+#ifdef CONFIG_CPU_ISOLATION
+	if (static_branch_unlikely(&housekeeping_overridden)) {
+		int cpu;
+
+		guard(preempt)();
+		if (!cpu_is_isolated(raw_smp_processor_id()))
+			return WORK_CPU_UNBOUND;
+
+		cpu = cpumask_any_and_distribute(housekeeping_cpumask(HK_TYPE_DOMAIN),
+						 cpu_online_mask);
+		if (cpu < nr_cpu_ids)
+			return cpu;
+	}
+#endif
+	return WORK_CPU_UNBOUND;
+}
+
 /*
  * It'd be great if the workqueue API had a way to pass
  * in a mask and had some smarts for more clever placement.
@@ -2265,7 +2284,7 @@ static int blk_mq_hctx_next_cpu(struct blk_mq_hw_ctx *hctx)
 
 	/* Switch to unbound if no allowable CPUs in this hctx */
 	if (hctx->queue->nr_hw_queues == 1 || blk_mq_hctx_empty_cpumask(hctx))
-		return WORK_CPU_UNBOUND;
+		return blk_mq_hctx_unbound_cpu();
 
 	if (--hctx->next_cpu_batch <= 0) {
 select_cpu:
@@ -2292,7 +2311,7 @@ static int blk_mq_hctx_next_cpu(struct blk_mq_hw_ctx *hctx)
 		 */
 		hctx->next_cpu = next_cpu;
 		hctx->next_cpu_batch = 1;
-		return WORK_CPU_UNBOUND;
+		return blk_mq_hctx_unbound_cpu();
 	}
 
 	hctx->next_cpu = next_cpu;
-- 
2.55.0


^ permalink raw reply related	[flat|nested] only message in thread

only message in thread, other threads:[~2026-11-21 23:52 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-21 23:52 [PATCH v5] blk-mq: dispatch kworker to housekeeping CPU when hctx cpumask is empty Aaron Tomlin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox