From: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
To: linux-kernel@vger.kernel.org
Cc: Peter Zijlstra <peterz@infradead.org>,
Ingo Molnar <mingo@redhat.com>,
rt@linutronix.de, tglx@linutronix.de,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
Jens Axboe <axboe@kernel.dk>
Subject: [PATCH 21/21] blk: mq: Convert to hotplug state machine
Date: Tue, 6 Sep 2016 19:04:57 +0200 [thread overview]
Message-ID: <20160906170457.32393-22-bigeasy@linutronix.de> (raw)
In-Reply-To: <20160906170457.32393-1-bigeasy@linutronix.de>
Install the callbacks via the state machine.
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
This patch against linux-next to take the block changes into account. It
depends on "blk: mq: reserve hotplug ID states for block"
block/blk-mq.c | 86 ++++++++++++++++++++++++++++-----------------------------
1 file changed, 43 insertions(+), 43 deletions(-)
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -2173,47 +2173,15 @@ static void blk_mq_queue_reinit(struct r
blk_mq_sysfs_register(q);
}
-static int blk_mq_queue_reinit_notify(struct notifier_block *nb,
- unsigned long action, void *hcpu)
+/*
+ * New online cpumask which is going to be set in this hotplug event.
+ * Declare this cpumasks as global as cpu-hotplug operation is invoked
+ * one-by-one and dynamically allocating this could result in a failure.
+ */
+static struct cpumask cpuhp_online_new;
+static int blk_mq_queue_reinit_work(void)
{
struct request_queue *q;
- int cpu = (unsigned long)hcpu;
- /*
- * New online cpumask which is going to be set in this hotplug event.
- * Declare this cpumasks as global as cpu-hotplug operation is invoked
- * one-by-one and dynamically allocating this could result in a failure.
- */
- static struct cpumask online_new;
-
- /*
- * Before hotadded cpu starts handling requests, new mappings must
- * be established. Otherwise, these requests in hw queue might
- * never be dispatched.
- *
- * For example, there is a single hw queue (hctx) and two CPU queues
- * (ctx0 for CPU0, and ctx1 for CPU1).
- *
- * Now CPU1 is just onlined and a request is inserted into
- * ctx1->rq_list and set bit0 in pending bitmap as ctx1->index_hw is
- * still zero.
- *
- * And then while running hw queue, flush_busy_ctxs() finds bit0 is
- * set in pending bitmap and tries to retrieve requests in
- * hctx->ctxs[0]->rq_list. But htx->ctxs[0] is a pointer to ctx0,
- * so the request in ctx1->rq_list is ignored.
- */
- switch (action & ~CPU_TASKS_FROZEN) {
- case CPU_DEAD:
- case CPU_UP_CANCELED:
- cpumask_copy(&online_new, cpu_online_mask);
- break;
- case CPU_UP_PREPARE:
- cpumask_copy(&online_new, cpu_online_mask);
- cpumask_set_cpu(cpu, &online_new);
- break;
- default:
- return NOTIFY_OK;
- }
mutex_lock(&all_q_mutex);
@@ -2237,13 +2205,44 @@ static int blk_mq_queue_reinit_notify(st
}
list_for_each_entry(q, &all_q_list, all_q_node)
- blk_mq_queue_reinit(q, &online_new);
+ blk_mq_queue_reinit(q, &cpuhp_online_new);
list_for_each_entry(q, &all_q_list, all_q_node)
blk_mq_unfreeze_queue(q);
mutex_unlock(&all_q_mutex);
- return NOTIFY_OK;
+ return 0;
+}
+
+static int blk_mq_queue_reinit_dead(unsigned int cpu)
+{
+ cpumask_copy(&cpuhp_online_new, cpu_online_mask);
+ return blk_mq_queue_reinit_work();
+}
+
+static int blk_mq_queue_reinit_prepare(unsigned int cpu)
+{
+ /*
+ * Before hotadded cpu starts handling requests, new mappings must
+ * be established. Otherwise, these requests in hw queue might
+ * never be dispatched.
+ *
+ * For example, there is a single hw queue (hctx) and two CPU queues
+ * (ctx0 for CPU0, and ctx1 for CPU1).
+ *
+ * Now CPU1 is just onlined and a request is inserted into
+ * ctx1->rq_list and set bit0 in pending bitmap as ctx1->index_hw is
+ * still zero.
+ *
+ * And then while running hw queue, flush_busy_ctxs() finds bit0 is
+ * set in pending bitmap and tries to retrieve requests in
+ * hctx->ctxs[0]->rq_list. But htx->ctxs[0] is a pointer to ctx0,
+ * so the request in ctx1->rq_list is ignored.
+ */
+
+ cpumask_copy(&cpuhp_online_new, cpu_online_mask);
+ cpumask_set_cpu(cpu, &cpuhp_online_new);
+ return blk_mq_queue_reinit_work();
}
static int __blk_mq_alloc_rq_maps(struct blk_mq_tag_set *set)
@@ -2447,8 +2446,9 @@ static int __init blk_mq_init(void)
{
blk_mq_cpu_init();
- hotcpu_notifier(blk_mq_queue_reinit_notify, 0);
-
+ cpuhp_setup_state_nocalls(CPUHP_BLK_MQ_PREPARE, "block/mq:prepare",
+ blk_mq_queue_reinit_prepare,
+ blk_mq_queue_reinit_dead);
return 0;
}
subsys_initcall(blk_mq_init);
prev parent reply other threads:[~2016-09-06 17:08 UTC|newest]
Thread overview: 66+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-06 17:04 cpu hotplug: convert more drivers (batch #3) Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 01/21] arm64: FP/SIMD: Convert to hotplug state machine Sebastian Andrzej Siewior
2016-09-06 18:14 ` Will Deacon
2016-09-19 19:51 ` [tip:smp/hotplug] arm64/FP/SIMD: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 02/21] ARM: shmobile: " Sebastian Andrzej Siewior
2016-09-06 18:05 ` Geert Uytterhoeven
2016-09-07 13:58 ` Sebastian Andrzej Siewior
2016-09-19 19:52 ` [tip:smp/hotplug] ARM/shmobile: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 03/21] ARM: OMAP: wakeupgen: " Sebastian Andrzej Siewior
2016-09-13 22:24 ` Tony Lindgren
2016-09-19 19:52 ` [tip:smp/hotplug] ARM/OMAP/wakeupgen: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 04/21] ia64: mca: " Sebastian Andrzej Siewior
2016-09-19 19:53 ` [tip:smp/hotplug] ia64/mca: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 05/21] sh: SH-X3 SMP: " Sebastian Andrzej Siewior
2016-09-19 19:53 ` [tip:smp/hotplug] sh/SH-X3 " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 06/21] x86: microcode: " Sebastian Andrzej Siewior
2016-09-07 11:36 ` Borislav Petkov
2016-09-07 14:51 ` Sebastian Andrzej Siewior
2016-09-07 16:02 ` Borislav Petkov
2016-09-07 16:45 ` [PATCH 06/21 v2] " Sebastian Andrzej Siewior
2016-09-19 19:54 ` [tip:smp/hotplug] x86/microcode: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 07/21] lib: irq_poll: " Sebastian Andrzej Siewior
2016-09-19 19:54 ` [tip:smp/hotplug] lib/irq_poll: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 08/21] block: softirq: " Sebastian Andrzej Siewior
2016-09-19 19:55 ` [tip:smp/hotplug] block/softirq: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 09/21] oprofile: timer: " Sebastian Andrzej Siewior
2016-09-19 19:55 ` [tip:smp/hotplug] oprofile/timer: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 10/21] virtio scsi: " Sebastian Andrzej Siewior
2016-09-19 19:56 ` [tip:smp/hotplug] " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 11/21] ACPI: processor: " Sebastian Andrzej Siewior
2016-09-06 21:08 ` Rafael J. Wysocki
2016-09-07 14:01 ` Sebastian Andrzej Siewior
2016-09-07 15:59 ` Rafael J. Wysocki
2016-09-19 19:56 ` [tip:smp/hotplug] ACPI/processor: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 12/21] cpufreq: " Sebastian Andrzej Siewior
2016-09-06 21:27 ` Rafael J. Wysocki
2016-09-07 14:18 ` Sebastian Andrzej Siewior
2016-09-07 15:58 ` Rafael J. Wysocki
2016-09-19 19:56 ` [tip:smp/hotplug] " tip-bot for Sebastian Andrzej Siewior
2016-09-20 14:56 ` [PATCH] cpufreq: fix up conversion " Sebastian Andrzej Siewior
2016-09-20 15:07 ` [tip:smp/hotplug] cpufreq: Fix " tip-bot for Sebastian Andrzej Siewior
2016-09-20 16:08 ` Borislav Petkov
2016-09-20 16:17 ` [PATCH] cpufreq: fix " Rafael J. Wysocki
2016-09-06 17:04 ` [PATCH 13/21] padata: Convert " Sebastian Andrzej Siewior
2016-09-19 19:57 ` [tip:smp/hotplug] " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 14/21] fault-injection: cpu: " Sebastian Andrzej Siewior
2016-09-19 19:57 ` [tip:smp/hotplug] fault-injection/cpu: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 15/21] mips: octeon: smp: " Sebastian Andrzej Siewior
2016-09-07 8:24 ` Matt Redfearn
2016-09-07 14:27 ` Sebastian Andrzej Siewior
2016-09-08 8:34 ` Matt Redfearn
2016-09-19 15:10 ` Ralf Baechle
2016-09-19 19:58 ` [tip:smp/hotplug] mips/octeon/smp: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 16/21] mips: loongson: smp: " Sebastian Andrzej Siewior
2016-09-19 15:10 ` Ralf Baechle
2016-09-19 19:58 ` [tip:smp/hotplug] mips/loongson/smp: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 17/21] s390: mm: pfault: " Sebastian Andrzej Siewior
2016-09-19 19:59 ` [tip:smp/hotplug] s390/mm/pfault: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 18/21] x86: apic: uv: " Sebastian Andrzej Siewior
2016-09-19 19:59 ` [tip:smp/hotplug] x86/apic/uv: " tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 19/21] blk: mq: reserve hotplug ID states for block Sebastian Andrzej Siewior
2016-09-19 20:03 ` [tip:smp/hotplug] blk/mq: Reserve hotplug ID states for block multiqueue tip-bot for Sebastian Andrzej Siewior
2016-09-19 21:24 ` [tip:smp/hotplug] blk/mq: Reserve hotplug " tip-bot for Sebastian Andrzej Siewior
2016-09-21 7:45 ` tip-bot for Sebastian Andrzej Siewior
2016-09-06 17:04 ` [PATCH 20/21] blk: mq: cpu-notif: Convert to hotplug state machine Sebastian Andrzej Siewior
2016-09-06 17:04 ` Sebastian Andrzej Siewior [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20160906170457.32393-22-bigeasy@linutronix.de \
--to=bigeasy@linutronix.de \
--cc=axboe@kernel.dk \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rt@linutronix.de \
--cc=tglx@linutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).