kernel_optimize_test/fs/btrfs/async-thread.c

335 lines
8.2 KiB
C
Raw Normal View History

/*
* Copyright (C) 2007 Oracle. All rights reserved.
* Copyright (C) 2014 Fujitsu. All rights reserved.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public
* License v2 as published by the Free Software Foundation.
*
* 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.
*
* You should have received a copy of the GNU General Public
* License along with this program; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
* Boston, MA 021110-1307, USA.
*/
#include <linux/kthread.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
#include <linux/slab.h>
#include <linux/list.h>
#include <linux/spinlock.h>
#include <linux/freezer.h>
#include <linux/workqueue.h>
#include "async-thread.h"
#include "ctree.h"
#define WORK_DONE_BIT 0
#define WORK_ORDER_DONE_BIT 1
#define WORK_HIGH_PRIO_BIT 2
Btrfs: Add ordered async work queues Btrfs uses kernel threads to create async work queues for cpu intensive operations such as checksumming and decompression. These work well, but they make it difficult to keep IO order intact. A single writepages call from pdflush or fsync will turn into a number of bios, and each bio is checksummed in parallel. Once the checksum is computed, the bio is sent down to the disk, and since we don't control the order in which the parallel operations happen, they might go down to the disk in almost any order. The code deals with this somewhat by having deep work queues for a single kernel thread, making it very likely that a single thread will process all the bios for a single inode. This patch introduces an explicitly ordered work queue. As work structs are placed into the queue they are put onto the tail of a list. They have three callbacks: ->func (cpu intensive processing here) ->ordered_func (order sensitive processing here) ->ordered_free (free the work struct, all processing is done) The work struct has three callbacks. The func callback does the cpu intensive work, and when it completes the work struct is marked as done. Every time a work struct completes, the list is checked to see if the head is marked as done. If so the ordered_func callback is used to do the order sensitive processing and the ordered_free callback is used to do any cleanup. Then we loop back and check the head of the list again. This patch also changes the checksumming code to use the ordered workqueues. One a 4 drive array, it increases streaming writes from 280MB/s to 350MB/s. Signed-off-by: Chris Mason <chris.mason@oracle.com>
2008-11-07 11:03:00 +08:00
#define NO_THRESHOLD (-1)
#define DFT_THRESHOLD (32)
struct __btrfs_workqueue {
struct workqueue_struct *normal_wq;
/* List head pointing to ordered work list */
struct list_head ordered_list;
/* Spinlock for ordered_list */
spinlock_t list_lock;
/* Thresholding related variants */
atomic_t pending;
int max_active;
int current_max;
int thresh;
unsigned int count;
spinlock_t thres_lock;
};
struct btrfs_workqueue {
struct __btrfs_workqueue *normal;
struct __btrfs_workqueue *high;
};
static inline struct __btrfs_workqueue
*__btrfs_alloc_workqueue(const char *name, int flags, int max_active,
int thresh)
{
struct __btrfs_workqueue *ret = kzalloc(sizeof(*ret), GFP_NOFS);
if (unlikely(!ret))
return NULL;
ret->max_active = max_active;
atomic_set(&ret->pending, 0);
if (thresh == 0)
thresh = DFT_THRESHOLD;
/* For low threshold, disabling threshold is a better choice */
if (thresh < DFT_THRESHOLD) {
ret->current_max = max_active;
ret->thresh = NO_THRESHOLD;
} else {
ret->current_max = 1;
ret->thresh = thresh;
}
if (flags & WQ_HIGHPRI)
ret->normal_wq = alloc_workqueue("%s-%s-high", flags,
ret->max_active,
"btrfs", name);
else
ret->normal_wq = alloc_workqueue("%s-%s", flags,
ret->max_active, "btrfs",
name);
if (unlikely(!ret->normal_wq)) {
kfree(ret);
return NULL;
}
INIT_LIST_HEAD(&ret->ordered_list);
spin_lock_init(&ret->list_lock);
spin_lock_init(&ret->thres_lock);
trace_btrfs_workqueue_alloc(ret, name, flags & WQ_HIGHPRI);
return ret;
}
static inline void
__btrfs_destroy_workqueue(struct __btrfs_workqueue *wq);
struct btrfs_workqueue *btrfs_alloc_workqueue(const char *name,
int flags,
int max_active,
int thresh)
{
struct btrfs_workqueue *ret = kzalloc(sizeof(*ret), GFP_NOFS);
if (unlikely(!ret))
return NULL;
ret->normal = __btrfs_alloc_workqueue(name, flags & ~WQ_HIGHPRI,
max_active, thresh);
if (unlikely(!ret->normal)) {
kfree(ret);
return NULL;
}
if (flags & WQ_HIGHPRI) {
ret->high = __btrfs_alloc_workqueue(name, flags, max_active,
thresh);
if (unlikely(!ret->high)) {
__btrfs_destroy_workqueue(ret->normal);
kfree(ret);
return NULL;
}
}
return ret;
}
/*
* Hook for threshold which will be called in btrfs_queue_work.
* This hook WILL be called in IRQ handler context,
* so workqueue_set_max_active MUST NOT be called in this hook
*/
static inline void thresh_queue_hook(struct __btrfs_workqueue *wq)
{
if (wq->thresh == NO_THRESHOLD)
return;
atomic_inc(&wq->pending);
}
/*
* Hook for threshold which will be called before executing the work,
* This hook is called in kthread content.
* So workqueue_set_max_active is called here.
*/
static inline void thresh_exec_hook(struct __btrfs_workqueue *wq)
{
int new_max_active;
long pending;
int need_change = 0;
if (wq->thresh == NO_THRESHOLD)
return;
atomic_dec(&wq->pending);
spin_lock(&wq->thres_lock);
/*
* Use wq->count to limit the calling frequency of
* workqueue_set_max_active.
*/
wq->count++;
wq->count %= (wq->thresh / 4);
if (!wq->count)
goto out;
new_max_active = wq->current_max;
/*
* pending may be changed later, but it's OK since we really
* don't need it so accurate to calculate new_max_active.
*/
pending = atomic_read(&wq->pending);
if (pending > wq->thresh)
new_max_active++;
if (pending < wq->thresh / 2)
new_max_active--;
new_max_active = clamp_val(new_max_active, 1, wq->max_active);
if (new_max_active != wq->current_max) {
need_change = 1;
wq->current_max = new_max_active;
}
out:
spin_unlock(&wq->thres_lock);
if (need_change) {
workqueue_set_max_active(wq->normal_wq, wq->current_max);
}
}
static void run_ordered_work(struct __btrfs_workqueue *wq)
{
struct list_head *list = &wq->ordered_list;
struct btrfs_work *work;
spinlock_t *lock = &wq->list_lock;
unsigned long flags;
while (1) {
spin_lock_irqsave(lock, flags);
if (list_empty(list))
break;
work = list_entry(list->next, struct btrfs_work,
ordered_list);
if (!test_bit(WORK_DONE_BIT, &work->flags))
break;
/*
* we are going to call the ordered done function, but
* we leave the work item on the list as a barrier so
* that later work items that are done don't have their
* functions called before this one returns
*/
if (test_and_set_bit(WORK_ORDER_DONE_BIT, &work->flags))
break;
trace_btrfs_ordered_sched(work);
spin_unlock_irqrestore(lock, flags);
work->ordered_func(work);
/* now take the lock again and drop our item from the list */
spin_lock_irqsave(lock, flags);
list_del(&work->ordered_list);
spin_unlock_irqrestore(lock, flags);
/*
* we don't want to call the ordered free functions
* with the lock held though
*/
work->ordered_free(work);
trace_btrfs_all_work_done(work);
}
spin_unlock_irqrestore(lock, flags);
}
static void normal_work_helper(struct work_struct *arg)
{
struct btrfs_work *work;
struct __btrfs_workqueue *wq;
int need_order = 0;
work = container_of(arg, struct btrfs_work, normal_work);
/*
* We should not touch things inside work in the following cases:
* 1) after work->func() if it has no ordered_free
* Since the struct is freed in work->func().
* 2) after setting WORK_DONE_BIT
* The work may be freed in other threads almost instantly.
* So we save the needed things here.
*/
if (work->ordered_func)
need_order = 1;
wq = work->wq;
trace_btrfs_work_sched(work);
thresh_exec_hook(wq);
work->func(work);
if (need_order) {
set_bit(WORK_DONE_BIT, &work->flags);
run_ordered_work(wq);
}
if (!need_order)
trace_btrfs_all_work_done(work);
}
void btrfs_init_work(struct btrfs_work *work,
btrfs_func_t func,
btrfs_func_t ordered_func,
btrfs_func_t ordered_free)
{
work->func = func;
work->ordered_func = ordered_func;
work->ordered_free = ordered_free;
INIT_WORK(&work->normal_work, normal_work_helper);
INIT_LIST_HEAD(&work->ordered_list);
work->flags = 0;
}
static inline void __btrfs_queue_work(struct __btrfs_workqueue *wq,
struct btrfs_work *work)
{
unsigned long flags;
work->wq = wq;
thresh_queue_hook(wq);
if (work->ordered_func) {
spin_lock_irqsave(&wq->list_lock, flags);
list_add_tail(&work->ordered_list, &wq->ordered_list);
spin_unlock_irqrestore(&wq->list_lock, flags);
}
queue_work(wq->normal_wq, &work->normal_work);
trace_btrfs_work_queued(work);
}
void btrfs_queue_work(struct btrfs_workqueue *wq,
struct btrfs_work *work)
{
struct __btrfs_workqueue *dest_wq;
if (test_bit(WORK_HIGH_PRIO_BIT, &work->flags) && wq->high)
dest_wq = wq->high;
else
dest_wq = wq->normal;
__btrfs_queue_work(dest_wq, work);
}
static inline void
__btrfs_destroy_workqueue(struct __btrfs_workqueue *wq)
{
destroy_workqueue(wq->normal_wq);
trace_btrfs_workqueue_destroy(wq);
kfree(wq);
}
void btrfs_destroy_workqueue(struct btrfs_workqueue *wq)
{
if (!wq)
return;
if (wq->high)
__btrfs_destroy_workqueue(wq->high);
__btrfs_destroy_workqueue(wq->normal);
kfree(wq);
}
void btrfs_workqueue_set_max(struct btrfs_workqueue *wq, int max)
{
wq->normal->max_active = max;
if (wq->high)
wq->high->max_active = max;
}
void btrfs_set_work_high_priority(struct btrfs_work *work)
{
set_bit(WORK_HIGH_PRIO_BIT, &work->flags);
}