From: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
To: linux-kernel@vger.kernel.org
Cc: tglx@linutronix.de, Nicholas Bellinger <nab@linux-iscsi.org>,
Andrew Morton <akpm@linux-foundation.org>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Subject: [PATCH] percpu_ida: Use _irqsave() instead of local_irq_save() + spin_lock
Date: Fri, 4 May 2018 17:32:18 +0200 [thread overview]
Message-ID: <20180504153218.7301-1-bigeasy@linutronix.de> (raw)
percpu_ida() decouples disabling interrupts from the locking operations.
This breaks some assumptions if the locking operations are replaced like
they are under -RT.
The same locking can be achieved by avoiding local_irq_save() and using
spin_lock_irqsave() instead. percpu_ida_alloc() gains one more
preemption point because after unlocking the fastpath and before the
pool lock is acquired, the interrupts are briefly enabled.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
lib/percpu_ida.c | 63 ++++++++++++++++--------------------------------
1 file changed, 21 insertions(+), 42 deletions(-)
diff --git a/lib/percpu_ida.c b/lib/percpu_ida.c
index 6016f1deb1f5..9bbd9c5d375a 100644
--- a/lib/percpu_ida.c
+++ b/lib/percpu_ida.c
@@ -112,18 +112,6 @@ static inline void alloc_global_tags(struct percpu_ida *pool,
min(pool->nr_free, pool->percpu_batch_size));
}
-static inline unsigned alloc_local_tag(struct percpu_ida_cpu *tags)
-{
- int tag = -ENOSPC;
-
- spin_lock(&tags->lock);
- if (tags->nr_free)
- tag = tags->freelist[--tags->nr_free];
- spin_unlock(&tags->lock);
-
- return tag;
-}
-
/**
* percpu_ida_alloc - allocate a tag
* @pool: pool to allocate from
@@ -147,20 +135,22 @@ int percpu_ida_alloc(struct percpu_ida *pool, int state)
DEFINE_WAIT(wait);
struct percpu_ida_cpu *tags;
unsigned long flags;
- int tag;
+ int tag = -ENOSPC;
- local_irq_save(flags);
- tags = this_cpu_ptr(pool->tag_cpu);
+ tags = raw_cpu_ptr(pool->tag_cpu);
+ spin_lock_irqsave(&tags->lock, flags);
/* Fastpath */
- tag = alloc_local_tag(tags);
- if (likely(tag >= 0)) {
- local_irq_restore(flags);
+ if (likely(tags->nr_free >= 0)) {
+ tag = tags->freelist[--tags->nr_free];
+ spin_unlock_irqrestore(&tags->lock, flags);
return tag;
}
+ spin_unlock_irqrestore(&tags->lock, flags);
while (1) {
- spin_lock(&pool->lock);
+ spin_lock_irqsave(&pool->lock, flags);
+ tags = this_cpu_ptr(pool->tag_cpu);
/*
* prepare_to_wait() must come before steal_tags(), in case
@@ -184,8 +174,7 @@ int percpu_ida_alloc(struct percpu_ida *pool, int state)
&pool->cpus_have_tags);
}
- spin_unlock(&pool->lock);
- local_irq_restore(flags);
+ spin_unlock_irqrestore(&pool->lock, flags);
if (tag >= 0 || state == TASK_RUNNING)
break;
@@ -196,9 +185,6 @@ int percpu_ida_alloc(struct percpu_ida *pool, int state)
}
schedule();
-
- local_irq_save(flags);
- tags = this_cpu_ptr(pool->tag_cpu);
}
if (state != TASK_RUNNING)
finish_wait(&pool->wait, &wait);
@@ -222,28 +208,24 @@ void percpu_ida_free(struct percpu_ida *pool, unsigned tag)
BUG_ON(tag >= pool->nr_tags);
- local_irq_save(flags);
- tags = this_cpu_ptr(pool->tag_cpu);
+ tags = raw_cpu_ptr(pool->tag_cpu);
- spin_lock(&tags->lock);
+ spin_lock_irqsave(&tags->lock, flags);
tags->freelist[tags->nr_free++] = tag;
nr_free = tags->nr_free;
- spin_unlock(&tags->lock);
if (nr_free == 1) {
cpumask_set_cpu(smp_processor_id(),
&pool->cpus_have_tags);
wake_up(&pool->wait);
}
+ spin_unlock_irqrestore(&tags->lock, flags);
if (nr_free == pool->percpu_max_size) {
- spin_lock(&pool->lock);
+ spin_lock_irqsave(&pool->lock, flags);
+ spin_lock(&tags->lock);
- /*
- * Global lock held and irqs disabled, don't need percpu
- * lock
- */
if (tags->nr_free == pool->percpu_max_size) {
move_tags(pool->freelist, &pool->nr_free,
tags->freelist, &tags->nr_free,
@@ -251,10 +233,9 @@ void percpu_ida_free(struct percpu_ida *pool, unsigned tag)
wake_up(&pool->wait);
}
- spin_unlock(&pool->lock);
+ spin_unlock(&tags->lock);
+ spin_unlock_irqrestore(&pool->lock, flags);
}
-
- local_irq_restore(flags);
}
EXPORT_SYMBOL_GPL(percpu_ida_free);
@@ -346,29 +327,27 @@ int percpu_ida_for_each_free(struct percpu_ida *pool, percpu_ida_cb fn,
struct percpu_ida_cpu *remote;
unsigned cpu, i, err = 0;
- local_irq_save(flags);
for_each_possible_cpu(cpu) {
remote = per_cpu_ptr(pool->tag_cpu, cpu);
- spin_lock(&remote->lock);
+ spin_lock_irqsave(&remote->lock, flags);
for (i = 0; i < remote->nr_free; i++) {
err = fn(remote->freelist[i], data);
if (err)
break;
}
- spin_unlock(&remote->lock);
+ spin_unlock_irqrestore(&remote->lock, flags);
if (err)
goto out;
}
- spin_lock(&pool->lock);
+ spin_lock_irqsave(&pool->lock, flags);
for (i = 0; i < pool->nr_free; i++) {
err = fn(pool->freelist[i], data);
if (err)
break;
}
- spin_unlock(&pool->lock);
+ spin_unlock_irqrestore(&pool->lock, flags);
out:
- local_irq_restore(flags);
return err;
}
EXPORT_SYMBOL_GPL(percpu_ida_for_each_free);
--
2.17.0
next reply other threads:[~2018-05-04 15:32 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-04 15:32 Sebastian Andrzej Siewior [this message]
2018-05-04 23:22 ` [PATCH] percpu_ida: Use _irqsave() instead of local_irq_save() + spin_lock Andrew Morton
2018-05-05 3:51 ` Matthew Wilcox
2018-05-05 14:10 ` Jens Axboe
2018-05-05 14:42 ` Jens Axboe
2018-05-05 15:52 ` Matthew Wilcox
2018-05-07 13:47 ` Matthew Wilcox
2018-05-07 21:34 ` Jens Axboe
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=20180504153218.7301-1-bigeasy@linutronix.de \
--to=bigeasy@linutronix.de \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=nab@linux-iscsi.org \
--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).