Linked list is used everywhere in the Linux kernel. However, if many
threads are trying to add or delete entries into the same linked list,
it can create a performance bottleneck.

This patch introduces a new per-cpu list subystem with associated
per-cpu locks for protecting each of the lists individually. This
allows list entries insertion and deletion operations to happen in
parallel instead of being serialized with a global list and lock.

List entry insertion is strictly per cpu. List deletion, however, can
happen in a cpu other than the one that did the insertion. So we still
need lock to protect the list. Because of that, there may still be
a small amount of contention when deletion is being done.

A new header file include/linux/percpu-list.h will be added with the
associated percpu_list structure. The following functions are used
to manage the per-cpu list:

 1. int init_percpu_list_head(struct percpu_list **pclist_handle)
 2. void percpu_list_add(struct percpu_list *new,
                         struct percpu_list *head)
 3. void percpu_list_del(struct percpu_list *entry)

Signed-off-by: Waiman Long <waiman.l...@hpe.com>
---
 include/linux/percpu-list.h |  117 +++++++++++++++++++++++++++++++++++++++++++
 lib/Makefile                |    2 +-
 lib/percpu-list.c           |   80 +++++++++++++++++++++++++++++
 3 files changed, 198 insertions(+), 1 deletions(-)
 create mode 100644 include/linux/percpu-list.h
 create mode 100644 lib/percpu-list.c

diff --git a/include/linux/percpu-list.h b/include/linux/percpu-list.h
new file mode 100644
index 0000000..94be520
--- /dev/null
+++ b/include/linux/percpu-list.h
@@ -0,0 +1,117 @@
+/*
+ * Per-cpu list
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * (C) Copyright 2016 Hewlett-Packard Enterprise Development LP
+ *
+ * Authors: Waiman Long <waiman.l...@hpe.com>
+ */
+#ifndef __LINUX_PERCPU_LIST_H
+#define __LINUX_PERCPU_LIST_H
+
+#include <linux/spinlock.h>
+#include <linux/list.h>
+#include <linux/percpu.h>
+
+/*
+ * include/linux/percpu-list.h
+ *
+ * A per-cpu list protected by a per-cpu spinlock.
+ *
+ * The list head percpu_list structure contains the spinlock, the other
+ * entries in the list contain the spinlock pointer.
+ */
+struct percpu_list {
+       struct list_head list;
+       union {
+               spinlock_t lock;        /* For list head */
+               spinlock_t *lockptr;    /* For other entries */
+       };
+};
+
+/*
+ * A simplified for_all_percpu_list_entries macro without the next and pchead
+ * parameters.
+ */
+#define for_all_percpu_list_entries_simple(pos, pclock, head, member)  \
+       for_all_percpu_list_entries(pos, next, pchead, pclock, head, member)
+
+#define PERCPU_LIST_HEAD_INIT(name)                            \
+       {                                                       \
+               .list.prev = &name.list,                        \
+               .list.next = &name.list,                        \
+               .list.lock = __SPIN_LOCK_UNLOCKED(name),        \
+       }
+
+#define PERCPU_LIST_ENTRY_INIT(name)           \
+       {                                       \
+               .list.prev = &name.list,        \
+               .list.next = &name.list,        \
+               .list.lockptr = NULL            \
+       }
+
+static inline void INIT_PERCPU_LIST_HEAD(struct percpu_list *pcpu_list)
+{
+       INIT_LIST_HEAD(&pcpu_list->list);
+       pcpu_list->lock = __SPIN_LOCK_UNLOCKED(&pcpu_list->lock);
+}
+
+static inline void INIT_PERCPU_LIST_ENTRY(struct percpu_list *pcpu_list)
+{
+       INIT_LIST_HEAD(&pcpu_list->list);
+       pcpu_list->lockptr = NULL;
+}
+
+#define PERCPU_LIST_HEAD(name) struct percpu_list __percpu *name
+
+static inline void free_percpu_list_head(struct percpu_list **pclist_handle)
+{
+       free_percpu(*pclist_handle);
+       *pclist_handle = NULL;
+}
+
+static inline bool percpu_list_empty(struct percpu_list *pcpu_list)
+{
+       int cpu;
+
+       for_each_possible_cpu (cpu)
+               if (!list_empty(&per_cpu_ptr(pcpu_list, cpu)->list))
+                       return false;
+       return true;
+}
+
+/**
+ * for_all_percpu_list_entries - iterate over all the per-cpu list with locking
+ * @pos:       the type * to use as a loop cursor for the current .
+ * @next:      an internal type * variable pointing to the next entry
+ * @pchead:    an internal struct list * of percpu list head
+ * @pclock:    an internal variable for the current per-cpu spinlock
+ * @head:      the head of the per-cpu list
+ * @member:    the name of the per-cpu list within the struct
+ */
+#define for_all_percpu_list_entries(pos, next, pchead, pclock, head, member)\
+       {                                                                \
+       int cpu;                                                         \
+       for_each_possible_cpu (cpu) {                                    \
+               typeof(*pos) *next;                                      \
+               spinlock_t *pclock = per_cpu_ptr(&(head)->lock, cpu);    \
+               struct list_head *pchead = &per_cpu_ptr(head, cpu)->list;\
+               spin_lock(pclock);                                       \
+               list_for_each_entry_safe(pos, next, pchead, member.list)
+
+#define end_all_percpu_list_entries(pclock)    spin_unlock(pclock); } }
+
+extern int init_percpu_list_head(struct percpu_list **pclist_handle);
+extern void percpu_list_add(struct percpu_list *new, struct percpu_list *head);
+extern void percpu_list_del(struct percpu_list *entry);
+
+#endif /* __LINUX_PERCPU_LIST_H */
diff --git a/lib/Makefile b/lib/Makefile
index a7c26a4..71a25d4 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -27,7 +27,7 @@ obj-y += bcd.o div64.o sort.o parser.o halfmd4.o 
debug_locks.o random32.o \
         gcd.o lcm.o list_sort.o uuid.o flex_array.o iov_iter.o clz_ctz.o \
         bsearch.o find_bit.o llist.o memweight.o kfifo.o \
         percpu-refcount.o percpu_ida.o rhashtable.o reciprocal_div.o \
-        once.o
+        once.o percpu-list.o
 obj-y += string_helpers.o
 obj-$(CONFIG_TEST_STRING_HELPERS) += test-string_helpers.o
 obj-y += hexdump.o
diff --git a/lib/percpu-list.c b/lib/percpu-list.c
new file mode 100644
index 0000000..e5c04bf
--- /dev/null
+++ b/lib/percpu-list.c
@@ -0,0 +1,80 @@
+/*
+ * Per-cpu list
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * (C) Copyright 2016 Hewlett-Packard Enterprise Development LP
+ *
+ * Authors: Waiman Long <waiman.l...@hpe.com>
+ */
+#include <linux/percpu-list.h>
+
+/*
+ * Initialize the per-cpu list
+ */
+int init_percpu_list_head(struct percpu_list **pclist_handle)
+{
+       struct percpu_list *pclist = alloc_percpu(struct percpu_list);
+       int cpu;
+
+       if (!pclist)
+               return -ENOMEM;
+
+       for_each_possible_cpu (cpu)
+               INIT_PERCPU_LIST_HEAD(per_cpu_ptr(pclist, cpu));
+
+       *pclist_handle = pclist;
+       return 0;
+}
+
+/*
+ * List selection is based on the CPU being used when the percpu_list_add()
+ * function is called. However, deletion may be done by a different CPU.
+ * So we still need to use a lock to protect the content of the list.
+ */
+void percpu_list_add(struct percpu_list *new, struct percpu_list *head)
+{
+       spinlock_t *lock;
+
+       /*
+        * We need to disable preemption before accessing the per-cpu data
+        * to make sure that the cpu won't be changed because of preemption.
+        */
+       preempt_disable();
+       lock = this_cpu_ptr(&head->lock);
+       spin_lock(lock);
+       new->lockptr = lock;
+       list_add(&new->list, this_cpu_ptr(&head->list));
+       spin_unlock(lock);
+       preempt_enable();
+}
+
+/*
+ * Delete an entry from a percpu list
+ *
+ * We need to check the lock pointer again after taking the lock to guard
+ * against concurrent delete of the same entry. If the lock pointer changes
+ * or becomes NULL, we assume that the deletion was done elsewhere.
+ */
+void percpu_list_del(struct percpu_list *entry)
+{
+       spinlock_t *lock = READ_ONCE(entry->lockptr);
+
+       if (unlikely(!lock))
+               return;
+
+       spin_lock(lock);
+       if (likely(entry->lockptr && (lock == entry->lockptr))) {
+               list_del_init(&entry->list);
+               entry->lockptr = NULL;
+       }
+       spin_unlock(lock);
+}
-- 
1.7.1

Reply via email to