[PATCH -mm 1/2] irq_work, Use llist in irq_work

From: Huang Ying
Date: Tue Aug 30 2011 - 01:16:58 EST


Use llist in irq_work instead of the lock-less linked list
implementation in irq_work to avoid the code duplication.

Signed-off-by: Huang Ying <ying.huang@xxxxxxxxx>
Cc: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
---
include/linux/irq_work.h | 15 ++++---
init/Kconfig | 1
kernel/irq_work.c | 92 ++++++++++++++++++-----------------------------
3 files changed, 47 insertions(+), 61 deletions(-)

--- a/include/linux/irq_work.h
+++ b/include/linux/irq_work.h
@@ -1,20 +1,23 @@
#ifndef _LINUX_IRQ_WORK_H
#define _LINUX_IRQ_WORK_H

+#include <linux/llist.h>
+
struct irq_work {
- struct irq_work *next;
+ unsigned long flags;
+ struct llist_node llnode;
void (*func)(struct irq_work *);
};

static inline
-void init_irq_work(struct irq_work *entry, void (*func)(struct irq_work *))
+void init_irq_work(struct irq_work *work, void (*func)(struct irq_work *))
{
- entry->next = NULL;
- entry->func = func;
+ work->flags = 0;
+ work->func = func;
}

-bool irq_work_queue(struct irq_work *entry);
+bool irq_work_queue(struct irq_work *work);
void irq_work_run(void);
-void irq_work_sync(struct irq_work *entry);
+void irq_work_sync(struct irq_work *work);

#endif /* _LINUX_IRQ_WORK_H */
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -26,6 +26,7 @@ config HAVE_IRQ_WORK
config IRQ_WORK
bool
depends on HAVE_IRQ_WORK
+ select LLIST

menu "General setup"

--- a/kernel/irq_work.c
+++ b/kernel/irq_work.c
@@ -19,49 +19,34 @@
* claimed NULL, 3 -> {pending} : claimed to be enqueued
* pending next, 3 -> {busy} : queued, pending callback
* busy NULL, 2 -> {free, claimed} : callback in progress, can be claimed
- *
- * We use the lower two bits of the next pointer to keep PENDING and BUSY
- * flags.
*/

#define IRQ_WORK_PENDING 1UL
#define IRQ_WORK_BUSY 2UL
#define IRQ_WORK_FLAGS 3UL

-static inline bool irq_work_is_set(struct irq_work *entry, int flags)
-{
- return (unsigned long)entry->next & flags;
-}
-
-static inline struct irq_work *irq_work_next(struct irq_work *entry)
-{
- unsigned long next = (unsigned long)entry->next;
- next &= ~IRQ_WORK_FLAGS;
- return (struct irq_work *)next;
-}
+#define LIST_NONEMPTY_BIT 0

-static inline struct irq_work *next_flags(struct irq_work *entry, int flags)
-{
- unsigned long next = (unsigned long)entry;
- next |= flags;
- return (struct irq_work *)next;
-}
+struct irq_work_list {
+ unsigned long flags;
+ struct llist_head llist;
+};

-static DEFINE_PER_CPU(struct irq_work *, irq_work_list);
+static DEFINE_PER_CPU(struct irq_work_list, irq_work_lists);

/*
* Claim the entry so that no one else will poke at it.
*/
-static bool irq_work_claim(struct irq_work *entry)
+static bool irq_work_claim(struct irq_work *work)
{
- struct irq_work *next, *nflags;
+ unsigned long flags, nflags;

do {
- next = entry->next;
- if ((unsigned long)next & IRQ_WORK_PENDING)
+ flags = work->flags;
+ if (flags & IRQ_WORK_PENDING)
return false;
- nflags = next_flags(next, IRQ_WORK_FLAGS);
- } while (cmpxchg(&entry->next, next, nflags) != next);
+ nflags = flags | IRQ_WORK_FLAGS;
+ } while (cmpxchg(&work->flags, flags, nflags) != flags);

return true;
}
@@ -77,23 +62,19 @@ void __weak arch_irq_work_raise(void)
/*
* Queue the entry and raise the IPI if needed.
*/
-static void __irq_work_queue(struct irq_work *entry)
+static void __irq_work_queue(struct irq_work *work)
{
- struct irq_work *next;
+ struct irq_work_list *irq_work_list;

- preempt_disable();
+ irq_work_list = &get_cpu_var(irq_work_lists);

- do {
- next = __this_cpu_read(irq_work_list);
- /* Can assign non-atomic because we keep the flags set. */
- entry->next = next_flags(next, IRQ_WORK_FLAGS);
- } while (this_cpu_cmpxchg(irq_work_list, next, entry) != next);
+ llist_add(&work->llnode, &irq_work_list->llist);

/* The list was empty, raise self-interrupt to start processing. */
- if (!irq_work_next(entry))
+ if (!test_and_set_bit(LIST_NONEMPTY_BIT, &irq_work_list->flags))
arch_irq_work_raise();

- preempt_enable();
+ put_cpu_var(irq_work_list);
}

/*
@@ -102,16 +83,16 @@ static void __irq_work_queue(struct irq_
*
* Can be re-enqueued while the callback is still in progress.
*/
-bool irq_work_queue(struct irq_work *entry)
+bool irq_work_queue(struct irq_work *work)
{
- if (!irq_work_claim(entry)) {
+ if (!irq_work_claim(work)) {
/*
* Already enqueued, can't do!
*/
return false;
}

- __irq_work_queue(entry);
+ __irq_work_queue(work);
return true;
}
EXPORT_SYMBOL_GPL(irq_work_queue);
@@ -122,34 +103,35 @@ EXPORT_SYMBOL_GPL(irq_work_queue);
*/
void irq_work_run(void)
{
- struct irq_work *list;
+ struct irq_work *work;
+ struct irq_work_list *irq_work_list;
+ struct llist_node *llnode;

- if (this_cpu_read(irq_work_list) == NULL)
+ irq_work_list = &__get_cpu_var(irq_work_lists);
+ if (llist_empty(&irq_work_list->llist))
return;

BUG_ON(!in_irq());
BUG_ON(!irqs_disabled());

- list = this_cpu_xchg(irq_work_list, NULL);
-
- while (list != NULL) {
- struct irq_work *entry = list;
+ clear_bit(LIST_NONEMPTY_BIT, &irq_work_list->flags);
+ llnode = llist_del_all(&irq_work_list->llist);
+ while (llnode != NULL) {
+ work = llist_entry(llnode, struct irq_work, llnode);

- list = irq_work_next(list);
+ llnode = llnode->next;

/*
- * Clear the PENDING bit, after this point the @entry
+ * Clear the PENDING bit, after this point the @work
* can be re-used.
*/
- entry->next = next_flags(NULL, IRQ_WORK_BUSY);
- entry->func(entry);
+ work->flags = IRQ_WORK_BUSY;
+ work->func(work);
/*
* Clear the BUSY bit and return to the free state if
* no-one else claimed it meanwhile.
*/
- (void)cmpxchg(&entry->next,
- next_flags(NULL, IRQ_WORK_BUSY),
- NULL);
+ (void)cmpxchg(&work->flags, IRQ_WORK_BUSY, 0);
}
}
EXPORT_SYMBOL_GPL(irq_work_run);
@@ -158,11 +140,11 @@ EXPORT_SYMBOL_GPL(irq_work_run);
* Synchronize against the irq_work @entry, ensures the entry is not
* currently in use.
*/
-void irq_work_sync(struct irq_work *entry)
+void irq_work_sync(struct irq_work *work)
{
WARN_ON_ONCE(irqs_disabled());

- while (irq_work_is_set(entry, IRQ_WORK_BUSY))
+ while (work->flags & IRQ_WORK_BUSY)
cpu_relax();
}
EXPORT_SYMBOL_GPL(irq_work_sync);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/