2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* Copyright (C) 1991, 1992 Linus Torvalds
|
|
|
|
* Copyright (C) 1994, Karl Keyte: Added support for disk statistics
|
|
|
|
* Elevator latency, (C) 2000 Andrea Arcangeli <andrea@suse.de> SuSE
|
|
|
|
* Queue request tables / lock, selectable elevator, Jens Axboe <axboe@suse.de>
|
2008-01-31 13:03:55 +01:00
|
|
|
* kernel-doc documentation started by NeilBrown <neilb@cse.unsw.edu.au>
|
|
|
|
* - July2000
|
2005-04-17 00:20:36 +02:00
|
|
|
* bio rewrite, highmem i/o, etc, Jens Axboe <axboe@suse.de> - may 2001
|
|
|
|
*/
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This handles all read/write requests to block devices
|
|
|
|
*/
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/backing-dev.h>
|
|
|
|
#include <linux/bio.h>
|
|
|
|
#include <linux/blkdev.h>
|
|
|
|
#include <linux/highmem.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/kernel_stat.h>
|
|
|
|
#include <linux/string.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/completion.h>
|
|
|
|
#include <linux/slab.h>
|
|
|
|
#include <linux/swap.h>
|
|
|
|
#include <linux/writeback.h>
|
2006-12-10 11:19:35 +01:00
|
|
|
#include <linux/task_io_accounting_ops.h>
|
2006-03-23 20:00:26 +01:00
|
|
|
#include <linux/blktrace_api.h>
|
2006-12-08 11:39:46 +01:00
|
|
|
#include <linux/fault-inject.h>
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-29 14:51:59 +01:00
|
|
|
#include "blk.h"
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
static int __make_request(struct request_queue *q, struct bio *bio);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* For the allocated request tables
|
|
|
|
*/
|
2008-02-18 13:45:51 +01:00
|
|
|
static struct kmem_cache *request_cachep;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* For queue allocation
|
|
|
|
*/
|
2008-01-31 13:03:55 +01:00
|
|
|
struct kmem_cache *blk_requestq_cachep;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Controlling structure to kblockd
|
|
|
|
*/
|
2006-01-09 16:02:34 +01:00
|
|
|
static struct workqueue_struct *kblockd_workqueue;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-29 13:54:41 +01:00
|
|
|
static void drive_stat_acct(struct request *rq, int new_io)
|
|
|
|
{
|
2008-05-07 10:15:46 +02:00
|
|
|
struct hd_struct *part;
|
2008-01-29 13:54:41 +01:00
|
|
|
int rw = rq_data_dir(rq);
|
2008-08-25 12:47:21 +02:00
|
|
|
int cpu;
|
2008-01-29 13:54:41 +01:00
|
|
|
|
|
|
|
if (!blk_fs_request(rq) || !rq->rq_disk)
|
|
|
|
return;
|
|
|
|
|
2008-08-25 12:56:14 +02:00
|
|
|
cpu = part_stat_lock();
|
2008-09-03 09:03:02 +02:00
|
|
|
part = disk_map_sector_rcu(rq->rq_disk, rq->sector);
|
2008-08-25 12:47:21 +02:00
|
|
|
|
2008-05-07 10:15:46 +02:00
|
|
|
if (!new_io)
|
2008-08-25 12:56:14 +02:00
|
|
|
part_stat_inc(cpu, part, merges[rw]);
|
2008-05-07 10:15:46 +02:00
|
|
|
else {
|
2008-08-25 12:56:14 +02:00
|
|
|
part_round_stats(cpu, part);
|
|
|
|
part_inc_in_flight(part);
|
2008-01-29 13:54:41 +01:00
|
|
|
}
|
2008-09-03 09:03:02 +02:00
|
|
|
|
2008-08-25 12:56:14 +02:00
|
|
|
part_stat_unlock();
|
2008-01-29 13:54:41 +01:00
|
|
|
}
|
|
|
|
|
2008-01-29 14:51:59 +01:00
|
|
|
void blk_queue_congestion_threshold(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
int nr;
|
|
|
|
|
|
|
|
nr = q->nr_requests - (q->nr_requests / 8) + 1;
|
|
|
|
if (nr > q->nr_requests)
|
|
|
|
nr = q->nr_requests;
|
|
|
|
q->nr_congestion_on = nr;
|
|
|
|
|
|
|
|
nr = q->nr_requests - (q->nr_requests / 8) - (q->nr_requests / 16) - 1;
|
|
|
|
if (nr < 1)
|
|
|
|
nr = 1;
|
|
|
|
q->nr_congestion_off = nr;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blk_get_backing_dev_info - get the address of a queue's backing_dev_info
|
|
|
|
* @bdev: device
|
|
|
|
*
|
|
|
|
* Locates the passed device's request queue and returns the address of its
|
|
|
|
* backing_dev_info
|
|
|
|
*
|
|
|
|
* Will return NULL if the request queue cannot be located.
|
|
|
|
*/
|
|
|
|
struct backing_dev_info *blk_get_backing_dev_info(struct block_device *bdev)
|
|
|
|
{
|
|
|
|
struct backing_dev_info *ret = NULL;
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q = bdev_get_queue(bdev);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
if (q)
|
|
|
|
ret = &q->backing_dev_info;
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_get_backing_dev_info);
|
|
|
|
|
2008-04-29 09:54:36 +02:00
|
|
|
void blk_rq_init(struct request_queue *q, struct request *rq)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2008-04-25 12:26:28 +02:00
|
|
|
memset(rq, 0, sizeof(*rq));
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
INIT_LIST_HEAD(&rq->queuelist);
|
2008-09-13 20:26:01 +02:00
|
|
|
rq->cpu = -1;
|
2008-02-08 12:41:03 +01:00
|
|
|
rq->q = q;
|
|
|
|
rq->sector = rq->hard_sector = (sector_t) -1;
|
2006-07-13 11:55:04 +02:00
|
|
|
INIT_HLIST_NODE(&rq->hash);
|
|
|
|
RB_CLEAR_NODE(&rq->rb_node);
|
2008-04-29 09:54:39 +02:00
|
|
|
rq->cmd = rq->__cmd;
|
2008-02-08 12:41:03 +01:00
|
|
|
rq->tag = -1;
|
2005-04-17 00:20:36 +02:00
|
|
|
rq->ref_count = 1;
|
|
|
|
}
|
2008-04-29 09:54:36 +02:00
|
|
|
EXPORT_SYMBOL(blk_rq_init);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-09-27 12:46:13 +02:00
|
|
|
static void req_bio_endio(struct request *rq, struct bio *bio,
|
|
|
|
unsigned int nbytes, int error)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q = rq->q;
|
2006-01-06 09:51:03 +01:00
|
|
|
|
2007-09-27 12:46:13 +02:00
|
|
|
if (&q->bar_rq != rq) {
|
|
|
|
if (error)
|
|
|
|
clear_bit(BIO_UPTODATE, &bio->bi_flags);
|
|
|
|
else if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
|
|
|
|
error = -EIO;
|
2006-01-06 09:51:03 +01:00
|
|
|
|
2007-09-27 12:46:13 +02:00
|
|
|
if (unlikely(nbytes > bio->bi_size)) {
|
2008-01-31 13:03:55 +01:00
|
|
|
printk(KERN_ERR "%s: want %u bytes done, %u left\n",
|
2008-05-01 13:35:17 +02:00
|
|
|
__func__, nbytes, bio->bi_size);
|
2007-09-27 12:46:13 +02:00
|
|
|
nbytes = bio->bi_size;
|
|
|
|
}
|
2006-01-06 09:51:03 +01:00
|
|
|
|
2007-09-27 12:46:13 +02:00
|
|
|
bio->bi_size -= nbytes;
|
|
|
|
bio->bi_sector += (nbytes >> 9);
|
2008-06-30 20:04:41 +02:00
|
|
|
|
|
|
|
if (bio_integrity(bio))
|
|
|
|
bio_integrity_advance(bio, nbytes);
|
|
|
|
|
2007-09-27 12:46:13 +02:00
|
|
|
if (bio->bi_size == 0)
|
2007-09-27 12:47:43 +02:00
|
|
|
bio_endio(bio, error);
|
2007-09-27 12:46:13 +02:00
|
|
|
} else {
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Okay, this is the barrier request in progress, just
|
|
|
|
* record the error;
|
|
|
|
*/
|
|
|
|
if (error && !q->orderr)
|
|
|
|
q->orderr = error;
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
void blk_dump_rq_flags(struct request *rq, char *msg)
|
|
|
|
{
|
|
|
|
int bit;
|
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
printk(KERN_INFO "%s: dev %s: type=%x, flags=%x\n", msg,
|
2006-08-10 08:44:47 +02:00
|
|
|
rq->rq_disk ? rq->rq_disk->disk_name : "?", rq->cmd_type,
|
|
|
|
rq->cmd_flags);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
printk(KERN_INFO " sector %llu, nr/cnr %lu/%u\n",
|
|
|
|
(unsigned long long)rq->sector,
|
|
|
|
rq->nr_sectors,
|
|
|
|
rq->current_nr_sectors);
|
|
|
|
printk(KERN_INFO " bio %p, biotail %p, buffer %p, data %p, len %u\n",
|
|
|
|
rq->bio, rq->biotail,
|
|
|
|
rq->buffer, rq->data,
|
|
|
|
rq->data_len);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2006-08-10 08:44:47 +02:00
|
|
|
if (blk_pc_request(rq)) {
|
2008-01-31 13:03:55 +01:00
|
|
|
printk(KERN_INFO " cdb: ");
|
2008-04-29 14:37:52 +02:00
|
|
|
for (bit = 0; bit < BLK_MAX_CDB; bit++)
|
2005-04-17 00:20:36 +02:00
|
|
|
printk("%02x ", rq->cmd[bit]);
|
|
|
|
printk("\n");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_dump_rq_flags);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* "plug" the device if there are no outstanding requests: this will
|
|
|
|
* force the transfer to start only after we have put all the requests
|
|
|
|
* on the list.
|
|
|
|
*
|
|
|
|
* This is called with interrupts off and no requests on the queue and
|
|
|
|
* with the queue lock held.
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
void blk_plug_device(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
WARN_ON(!irqs_disabled());
|
|
|
|
|
|
|
|
/*
|
|
|
|
* don't plug a stopped queue, it must be paired with blk_start_queue()
|
|
|
|
* which will restart the queueing
|
|
|
|
*/
|
2006-04-19 10:14:49 +02:00
|
|
|
if (blk_queue_stopped(q))
|
2005-04-17 00:20:36 +02:00
|
|
|
return;
|
|
|
|
|
2008-07-03 13:18:54 +02:00
|
|
|
if (!queue_flag_test_and_set(QUEUE_FLAG_PLUGGED, q)) {
|
2005-04-17 00:20:36 +02:00
|
|
|
mod_timer(&q->unplug_timer, jiffies + q->unplug_delay);
|
2006-03-23 20:00:26 +01:00
|
|
|
blk_add_trace_generic(q, NULL, 0, BLK_TA_PLUG);
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_plug_device);
|
|
|
|
|
2008-08-01 20:31:32 +02:00
|
|
|
/**
|
|
|
|
* blk_plug_device_unlocked - plug a device without queue lock held
|
|
|
|
* @q: The &struct request_queue to plug
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Like @blk_plug_device(), but grabs the queue lock and disables
|
|
|
|
* interrupts.
|
|
|
|
**/
|
|
|
|
void blk_plug_device_unlocked(struct request_queue *q)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(q->queue_lock, flags);
|
|
|
|
blk_plug_device(q);
|
|
|
|
spin_unlock_irqrestore(q->queue_lock, flags);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_plug_device_unlocked);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* remove the queue from the plugged list, if present. called with
|
|
|
|
* queue lock held and interrupts disabled.
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
int blk_remove_plug(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
WARN_ON(!irqs_disabled());
|
|
|
|
|
2008-07-03 13:18:54 +02:00
|
|
|
if (!queue_flag_test_and_clear(QUEUE_FLAG_PLUGGED, q))
|
2005-04-17 00:20:36 +02:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
del_timer(&q->unplug_timer);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_remove_plug);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* remove the plug and let it rip..
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
void __generic_unplug_device(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2006-04-19 10:14:49 +02:00
|
|
|
if (unlikely(blk_queue_stopped(q)))
|
2005-04-17 00:20:36 +02:00
|
|
|
return;
|
|
|
|
|
|
|
|
if (!blk_remove_plug(q))
|
|
|
|
return;
|
|
|
|
|
2005-06-27 10:55:12 +02:00
|
|
|
q->request_fn(q);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(__generic_unplug_device);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* generic_unplug_device - fire a request queue
|
2007-07-24 09:28:11 +02:00
|
|
|
* @q: The &struct request_queue in question
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Linux uses plugging to build bigger requests queues before letting
|
|
|
|
* the device have at them. If a queue is plugged, the I/O scheduler
|
|
|
|
* is still adding and merging requests on the queue. Once the queue
|
|
|
|
* gets unplugged, the request_fn defined for the queue is invoked and
|
|
|
|
* transfers started.
|
|
|
|
**/
|
2007-07-24 09:28:11 +02:00
|
|
|
void generic_unplug_device(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2008-05-07 09:48:17 +02:00
|
|
|
if (blk_queue_plugged(q)) {
|
|
|
|
spin_lock_irq(q->queue_lock);
|
|
|
|
__generic_unplug_device(q);
|
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(generic_unplug_device);
|
|
|
|
|
|
|
|
static void blk_backing_dev_unplug(struct backing_dev_info *bdi,
|
|
|
|
struct page *page)
|
|
|
|
{
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q = bdi->unplug_io_data;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-11-07 20:26:56 +01:00
|
|
|
blk_unplug(q);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2008-01-29 14:53:40 +01:00
|
|
|
void blk_unplug_work(struct work_struct *work)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q =
|
|
|
|
container_of(work, struct request_queue, unplug_work);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2006-03-23 20:00:26 +01:00
|
|
|
blk_add_trace_pdu_int(q, BLK_TA_UNPLUG_IO, NULL,
|
|
|
|
q->rq.count[READ] + q->rq.count[WRITE]);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
q->unplug_fn(q);
|
|
|
|
}
|
|
|
|
|
2008-01-29 14:53:40 +01:00
|
|
|
void blk_unplug_timeout(unsigned long data)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q = (struct request_queue *)data;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2006-03-23 20:00:26 +01:00
|
|
|
blk_add_trace_pdu_int(q, BLK_TA_UNPLUG_TIMER, NULL,
|
|
|
|
q->rq.count[READ] + q->rq.count[WRITE]);
|
|
|
|
|
2008-07-28 13:08:45 +02:00
|
|
|
kblockd_schedule_work(q, &q->unplug_work);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2007-11-07 20:26:56 +01:00
|
|
|
void blk_unplug(struct request_queue *q)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* devices don't necessarily have an ->unplug_fn defined
|
|
|
|
*/
|
|
|
|
if (q->unplug_fn) {
|
|
|
|
blk_add_trace_pdu_int(q, BLK_TA_UNPLUG_IO, NULL,
|
|
|
|
q->rq.count[READ] + q->rq.count[WRITE]);
|
|
|
|
|
|
|
|
q->unplug_fn(q);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_unplug);
|
|
|
|
|
2008-09-13 20:26:01 +02:00
|
|
|
static void blk_invoke_request_fn(struct request_queue *q)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* one level of recursion is ok and is much faster than kicking
|
|
|
|
* the unplug handling
|
|
|
|
*/
|
|
|
|
if (!queue_flag_test_and_set(QUEUE_FLAG_REENTER, q)) {
|
|
|
|
q->request_fn(q);
|
|
|
|
queue_flag_clear(QUEUE_FLAG_REENTER, q);
|
|
|
|
} else {
|
|
|
|
queue_flag_set(QUEUE_FLAG_PLUGGED, q);
|
|
|
|
kblockd_schedule_work(q, &q->unplug_work);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/**
|
|
|
|
* blk_start_queue - restart a previously stopped queue
|
2007-07-24 09:28:11 +02:00
|
|
|
* @q: The &struct request_queue in question
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* blk_start_queue() will clear the stop flag on the queue, and call
|
|
|
|
* the request_fn for the queue if it was in a stopped state when
|
|
|
|
* entered. Also see blk_stop_queue(). Queue lock must be held.
|
|
|
|
**/
|
2007-07-24 09:28:11 +02:00
|
|
|
void blk_start_queue(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2006-06-05 12:09:01 +02:00
|
|
|
WARN_ON(!irqs_disabled());
|
|
|
|
|
2008-04-29 14:48:33 +02:00
|
|
|
queue_flag_clear(QUEUE_FLAG_STOPPED, q);
|
2008-09-13 20:26:01 +02:00
|
|
|
blk_invoke_request_fn(q);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_start_queue);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blk_stop_queue - stop a queue
|
2007-07-24 09:28:11 +02:00
|
|
|
* @q: The &struct request_queue in question
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* The Linux block layer assumes that a block driver will consume all
|
|
|
|
* entries on the request queue when the request_fn strategy is called.
|
|
|
|
* Often this will not happen, because of hardware limitations (queue
|
|
|
|
* depth settings). If a device driver gets a 'queue full' response,
|
|
|
|
* or if it simply chooses not to queue more I/O at one point, it can
|
|
|
|
* call this function to prevent the request_fn from being called until
|
|
|
|
* the driver has signalled it's ready to go again. This happens by calling
|
|
|
|
* blk_start_queue() to restart queue operations. Queue lock must be held.
|
|
|
|
**/
|
2007-07-24 09:28:11 +02:00
|
|
|
void blk_stop_queue(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
blk_remove_plug(q);
|
2008-04-29 14:48:33 +02:00
|
|
|
queue_flag_set(QUEUE_FLAG_STOPPED, q);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_stop_queue);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blk_sync_queue - cancel any pending callbacks on a queue
|
|
|
|
* @q: the queue
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* The block layer may perform asynchronous callback activity
|
|
|
|
* on a queue, such as calling the unplug function after a timeout.
|
|
|
|
* A block device may call blk_sync_queue to ensure that any
|
|
|
|
* such activity is cancelled, thus allowing it to release resources
|
2007-05-09 08:57:56 +02:00
|
|
|
* that the callbacks might use. The caller must already have made sure
|
2005-04-17 00:20:36 +02:00
|
|
|
* that its ->make_request_fn will not re-add plugging prior to calling
|
|
|
|
* this function.
|
|
|
|
*
|
|
|
|
*/
|
|
|
|
void blk_sync_queue(struct request_queue *q)
|
|
|
|
{
|
|
|
|
del_timer_sync(&q->unplug_timer);
|
2007-10-23 15:08:19 +02:00
|
|
|
kblockd_flush_work(&q->unplug_work);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_sync_queue);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blk_run_queue - run a single device queue
|
|
|
|
* @q: The queue to run
|
|
|
|
*/
|
2008-04-29 14:48:33 +02:00
|
|
|
void __blk_run_queue(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
blk_remove_plug(q);
|
2006-05-11 08:20:16 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Only recurse once to avoid overrunning the stack, let the unplug
|
|
|
|
* handling reinvoke the handler shortly if we already got there.
|
|
|
|
*/
|
2008-09-13 20:26:01 +02:00
|
|
|
if (!elv_queue_empty(q))
|
|
|
|
blk_invoke_request_fn(q);
|
2008-04-29 14:48:33 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(__blk_run_queue);
|
2006-05-11 08:20:16 +02:00
|
|
|
|
2008-04-29 14:48:33 +02:00
|
|
|
/**
|
|
|
|
* blk_run_queue - run a single device queue
|
|
|
|
* @q: The queue to run
|
|
|
|
*/
|
|
|
|
void blk_run_queue(struct request_queue *q)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(q->queue_lock, flags);
|
|
|
|
__blk_run_queue(q);
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_unlock_irqrestore(q->queue_lock, flags);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_run_queue);
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
void blk_put_queue(struct request_queue *q)
|
2006-03-19 00:34:37 +01:00
|
|
|
{
|
|
|
|
kobject_put(&q->kobj);
|
|
|
|
}
|
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
void blk_cleanup_queue(struct request_queue *q)
|
2006-03-19 00:34:37 +01:00
|
|
|
{
|
|
|
|
mutex_lock(&q->sysfs_lock);
|
2008-04-29 14:48:33 +02:00
|
|
|
queue_flag_set_unlocked(QUEUE_FLAG_DEAD, q);
|
2006-03-19 00:34:37 +01:00
|
|
|
mutex_unlock(&q->sysfs_lock);
|
|
|
|
|
|
|
|
if (q->elevator)
|
|
|
|
elevator_exit(q->elevator);
|
|
|
|
|
|
|
|
blk_put_queue(q);
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
EXPORT_SYMBOL(blk_cleanup_queue);
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
static int blk_init_free_list(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct request_list *rl = &q->rq;
|
|
|
|
|
|
|
|
rl->count[READ] = rl->count[WRITE] = 0;
|
|
|
|
rl->starved[READ] = rl->starved[WRITE] = 0;
|
2005-10-28 08:29:39 +02:00
|
|
|
rl->elvpriv = 0;
|
2005-04-17 00:20:36 +02:00
|
|
|
init_waitqueue_head(&rl->wait[READ]);
|
|
|
|
init_waitqueue_head(&rl->wait[WRITE]);
|
|
|
|
|
2005-06-23 09:08:19 +02:00
|
|
|
rl->rq_pool = mempool_create_node(BLKDEV_MIN_RQ, mempool_alloc_slab,
|
|
|
|
mempool_free_slab, request_cachep, q->node);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
if (!rl->rq_pool)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *blk_alloc_queue(gfp_t gfp_mask)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2005-06-23 09:08:19 +02:00
|
|
|
return blk_alloc_queue_node(gfp_mask, -1);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_alloc_queue);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id)
|
2005-06-23 09:08:19 +02:00
|
|
|
{
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q;
|
2007-10-17 08:25:46 +02:00
|
|
|
int err;
|
2005-06-23 09:08:19 +02:00
|
|
|
|
2008-01-29 14:51:59 +01:00
|
|
|
q = kmem_cache_alloc_node(blk_requestq_cachep,
|
2007-07-17 13:03:29 +02:00
|
|
|
gfp_mask | __GFP_ZERO, node_id);
|
2005-04-17 00:20:36 +02:00
|
|
|
if (!q)
|
|
|
|
return NULL;
|
|
|
|
|
2007-10-17 08:25:46 +02:00
|
|
|
q->backing_dev_info.unplug_io_fn = blk_backing_dev_unplug;
|
|
|
|
q->backing_dev_info.unplug_io_data = q;
|
|
|
|
err = bdi_init(&q->backing_dev_info);
|
|
|
|
if (err) {
|
2008-01-29 14:51:59 +01:00
|
|
|
kmem_cache_free(blk_requestq_cachep, q);
|
2007-10-17 08:25:46 +02:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
init_timer(&q->unplug_timer);
|
2006-03-19 00:34:37 +01:00
|
|
|
|
2008-01-29 14:51:59 +01:00
|
|
|
kobject_init(&q->kobj, &blk_queue_ktype);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2006-03-19 00:34:37 +01:00
|
|
|
mutex_init(&q->sysfs_lock);
|
2008-05-15 01:05:54 +02:00
|
|
|
spin_lock_init(&q->__queue_lock);
|
2006-03-19 00:34:37 +01:00
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
return q;
|
|
|
|
}
|
2005-06-23 09:08:19 +02:00
|
|
|
EXPORT_SYMBOL(blk_alloc_queue_node);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/**
|
|
|
|
* blk_init_queue - prepare a request queue for use with a block device
|
|
|
|
* @rfn: The function to be called to process requests that have been
|
|
|
|
* placed on the queue.
|
|
|
|
* @lock: Request queue spin lock
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* If a block device wishes to use the standard request handling procedures,
|
|
|
|
* which sorts requests and coalesces adjacent requests, then it must
|
|
|
|
* call blk_init_queue(). The function @rfn will be called when there
|
|
|
|
* are requests on the queue that need to be processed. If the device
|
|
|
|
* supports plugging, then @rfn may not be called immediately when requests
|
|
|
|
* are available on the queue, but may be called at some time later instead.
|
|
|
|
* Plugged queues are generally unplugged when a buffer belonging to one
|
|
|
|
* of the requests on the queue is needed, or due to memory pressure.
|
|
|
|
*
|
|
|
|
* @rfn is not required, or even expected, to remove all requests off the
|
|
|
|
* queue, but only as many as it can handle at a time. If it does leave
|
|
|
|
* requests on the queue, it is responsible for arranging that the requests
|
|
|
|
* get dealt with eventually.
|
|
|
|
*
|
|
|
|
* The queue spin lock must be held while manipulating the requests on the
|
2006-06-05 12:09:01 +02:00
|
|
|
* request queue; this lock will be taken also from interrupt context, so irq
|
|
|
|
* disabling is needed for it.
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
2008-08-19 20:13:11 +02:00
|
|
|
* Function returns a pointer to the initialized request queue, or %NULL if
|
2005-04-17 00:20:36 +02:00
|
|
|
* it didn't succeed.
|
|
|
|
*
|
|
|
|
* Note:
|
|
|
|
* blk_init_queue() must be paired with a blk_cleanup_queue() call
|
|
|
|
* when the block device is deactivated (such as at module unload).
|
|
|
|
**/
|
2005-06-23 09:08:19 +02:00
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *blk_init_queue(request_fn_proc *rfn, spinlock_t *lock)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2005-06-23 09:08:19 +02:00
|
|
|
return blk_init_queue_node(rfn, lock, -1);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_init_queue);
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *
|
2005-06-23 09:08:19 +02:00
|
|
|
blk_init_queue_node(request_fn_proc *rfn, spinlock_t *lock, int node_id)
|
|
|
|
{
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q = blk_alloc_queue_node(GFP_KERNEL, node_id);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
if (!q)
|
|
|
|
return NULL;
|
|
|
|
|
2005-06-23 09:08:19 +02:00
|
|
|
q->node = node_id;
|
2006-03-18 19:50:00 +01:00
|
|
|
if (blk_init_free_list(q)) {
|
2008-01-29 14:51:59 +01:00
|
|
|
kmem_cache_free(blk_requestq_cachep, q);
|
2006-03-18 19:50:00 +01:00
|
|
|
return NULL;
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2005-04-12 23:22:06 +02:00
|
|
|
/*
|
|
|
|
* if caller didn't supply a lock, they get per-queue locking with
|
|
|
|
* our embedded lock
|
|
|
|
*/
|
2008-05-15 01:05:54 +02:00
|
|
|
if (!lock)
|
2005-04-12 23:22:06 +02:00
|
|
|
lock = &q->__queue_lock;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
q->request_fn = rfn;
|
|
|
|
q->prep_rq_fn = NULL;
|
|
|
|
q->unplug_fn = generic_unplug_device;
|
|
|
|
q->queue_flags = (1 << QUEUE_FLAG_CLUSTER);
|
|
|
|
q->queue_lock = lock;
|
|
|
|
|
|
|
|
blk_queue_segment_boundary(q, 0xffffffff);
|
|
|
|
|
|
|
|
blk_queue_make_request(q, __make_request);
|
|
|
|
blk_queue_max_segment_size(q, MAX_SEGMENT_SIZE);
|
|
|
|
|
|
|
|
blk_queue_max_hw_segments(q, MAX_HW_SEGMENTS);
|
|
|
|
blk_queue_max_phys_segments(q, MAX_PHYS_SEGMENTS);
|
|
|
|
|
2007-02-20 17:01:57 +01:00
|
|
|
q->sg_reserved_size = INT_MAX;
|
|
|
|
|
2008-08-16 07:10:05 +02:00
|
|
|
blk_set_cmd_filter_defaults(&q->cmd_filter);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* all done
|
|
|
|
*/
|
|
|
|
if (!elevator_init(q, NULL)) {
|
|
|
|
blk_queue_congestion_threshold(q);
|
|
|
|
return q;
|
|
|
|
}
|
|
|
|
|
2006-03-18 19:50:00 +01:00
|
|
|
blk_put_queue(q);
|
2005-04-17 00:20:36 +02:00
|
|
|
return NULL;
|
|
|
|
}
|
2005-06-23 09:08:19 +02:00
|
|
|
EXPORT_SYMBOL(blk_init_queue_node);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
int blk_get_queue(struct request_queue *q)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2005-06-23 09:08:53 +02:00
|
|
|
if (likely(!test_bit(QUEUE_FLAG_DEAD, &q->queue_flags))) {
|
2006-03-19 00:34:37 +01:00
|
|
|
kobject_get(&q->kobj);
|
2005-04-17 00:20:36 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
static inline void blk_free_request(struct request_queue *q, struct request *rq)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2006-08-10 08:44:47 +02:00
|
|
|
if (rq->cmd_flags & REQ_ELVPRIV)
|
2005-10-28 08:29:39 +02:00
|
|
|
elv_put_request(q, rq);
|
2005-04-17 00:20:36 +02:00
|
|
|
mempool_free(rq, q->rq.rq_pool);
|
|
|
|
}
|
|
|
|
|
2006-07-18 22:24:11 +02:00
|
|
|
static struct request *
|
2007-07-24 09:28:11 +02:00
|
|
|
blk_alloc_request(struct request_queue *q, int rw, int priv, gfp_t gfp_mask)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct request *rq = mempool_alloc(q->rq.rq_pool, gfp_mask);
|
|
|
|
|
|
|
|
if (!rq)
|
|
|
|
return NULL;
|
|
|
|
|
2008-04-29 09:54:36 +02:00
|
|
|
blk_rq_init(q, rq);
|
2008-04-25 12:26:28 +02:00
|
|
|
|
2006-08-10 08:59:11 +02:00
|
|
|
rq->cmd_flags = rw | REQ_ALLOCED;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2005-10-28 08:29:39 +02:00
|
|
|
if (priv) {
|
2006-07-28 09:32:57 +02:00
|
|
|
if (unlikely(elv_set_request(q, rq, gfp_mask))) {
|
2005-10-28 08:29:39 +02:00
|
|
|
mempool_free(rq, q->rq.rq_pool);
|
|
|
|
return NULL;
|
|
|
|
}
|
2006-08-10 08:44:47 +02:00
|
|
|
rq->cmd_flags |= REQ_ELVPRIV;
|
2005-10-28 08:29:39 +02:00
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2005-10-28 08:29:39 +02:00
|
|
|
return rq;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ioc_batching returns true if the ioc is a valid batching request and
|
|
|
|
* should be given priority access to a request.
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
static inline int ioc_batching(struct request_queue *q, struct io_context *ioc)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
if (!ioc)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Make sure the process is able to allocate at least 1 request
|
|
|
|
* even if the batch times out, otherwise we could theoretically
|
|
|
|
* lose wakeups.
|
|
|
|
*/
|
|
|
|
return ioc->nr_batch_requests == q->nr_batching ||
|
|
|
|
(ioc->nr_batch_requests > 0
|
|
|
|
&& time_before(jiffies, ioc->last_waited + BLK_BATCH_TIME));
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* ioc_set_batching sets ioc to be a new "batcher" if it is not one. This
|
|
|
|
* will cause the process to be a "batcher" on all queues in the system. This
|
|
|
|
* is the behaviour we want though - once it gets a wakeup it should be given
|
|
|
|
* a nice run.
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
static void ioc_set_batching(struct request_queue *q, struct io_context *ioc)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
if (!ioc || ioc_batching(q, ioc))
|
|
|
|
return;
|
|
|
|
|
|
|
|
ioc->nr_batch_requests = q->nr_batching;
|
|
|
|
ioc->last_waited = jiffies;
|
|
|
|
}
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
static void __freed_request(struct request_queue *q, int rw)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct request_list *rl = &q->rq;
|
|
|
|
|
|
|
|
if (rl->count[rw] < queue_congestion_off_threshold(q))
|
2006-10-20 08:28:15 +02:00
|
|
|
blk_clear_queue_congested(q, rw);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
if (rl->count[rw] + 1 <= q->nr_requests) {
|
|
|
|
if (waitqueue_active(&rl->wait[rw]))
|
|
|
|
wake_up(&rl->wait[rw]);
|
|
|
|
|
|
|
|
blk_clear_queue_full(q, rw);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* A request has just been released. Account for it, update the full and
|
|
|
|
* congestion status, wake up any waiters. Called under q->queue_lock.
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
static void freed_request(struct request_queue *q, int rw, int priv)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct request_list *rl = &q->rq;
|
|
|
|
|
|
|
|
rl->count[rw]--;
|
2005-10-28 08:29:39 +02:00
|
|
|
if (priv)
|
|
|
|
rl->elvpriv--;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
__freed_request(q, rw);
|
|
|
|
|
|
|
|
if (unlikely(rl->starved[rw ^ 1]))
|
|
|
|
__freed_request(q, rw ^ 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
#define blkdev_free_rq(list) list_entry((list)->next, struct request, queuelist)
|
|
|
|
/*
|
2005-06-29 05:45:14 +02:00
|
|
|
* Get a free request, queue_lock must be held.
|
|
|
|
* Returns NULL on failure, with queue_lock held.
|
|
|
|
* Returns !NULL on success, with queue_lock *not held*.
|
2005-04-17 00:20:36 +02:00
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
static struct request *get_request(struct request_queue *q, int rw_flags,
|
2006-12-13 13:02:26 +01:00
|
|
|
struct bio *bio, gfp_t gfp_mask)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct request *rq = NULL;
|
|
|
|
struct request_list *rl = &q->rq;
|
2005-11-12 11:09:12 +01:00
|
|
|
struct io_context *ioc = NULL;
|
2006-12-13 13:02:26 +01:00
|
|
|
const int rw = rw_flags & 0x01;
|
2005-11-12 11:09:12 +01:00
|
|
|
int may_queue, priv;
|
|
|
|
|
2006-12-13 13:02:26 +01:00
|
|
|
may_queue = elv_may_queue(q, rw_flags);
|
2005-11-12 11:09:12 +01:00
|
|
|
if (may_queue == ELV_MQUEUE_NO)
|
|
|
|
goto rq_starved;
|
|
|
|
|
|
|
|
if (rl->count[rw]+1 >= queue_congestion_on_threshold(q)) {
|
|
|
|
if (rl->count[rw]+1 >= q->nr_requests) {
|
2006-07-19 23:39:40 +02:00
|
|
|
ioc = current_io_context(GFP_ATOMIC, q->node);
|
2005-11-12 11:09:12 +01:00
|
|
|
/*
|
|
|
|
* The queue will fill after this allocation, so set
|
|
|
|
* it as full, and mark this process as "batching".
|
|
|
|
* This process will be allowed to complete a batch of
|
|
|
|
* requests, others will be blocked.
|
|
|
|
*/
|
|
|
|
if (!blk_queue_full(q, rw)) {
|
|
|
|
ioc_set_batching(q, ioc);
|
|
|
|
blk_set_queue_full(q, rw);
|
|
|
|
} else {
|
|
|
|
if (may_queue != ELV_MQUEUE_MUST
|
|
|
|
&& !ioc_batching(q, ioc)) {
|
|
|
|
/*
|
|
|
|
* The queue is full and the allocating
|
|
|
|
* process is not a "batcher", and not
|
|
|
|
* exempted by the IO scheduler
|
|
|
|
*/
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
2006-10-20 08:28:15 +02:00
|
|
|
blk_set_queue_congested(q, rw);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2005-06-28 16:35:11 +02:00
|
|
|
/*
|
|
|
|
* Only allow batching queuers to allocate up to 50% over the defined
|
|
|
|
* limit of requests, otherwise we could have thousands of requests
|
|
|
|
* allocated with any setting of ->nr_requests
|
|
|
|
*/
|
2005-06-29 16:15:40 +02:00
|
|
|
if (rl->count[rw] >= (3 * q->nr_requests / 2))
|
2005-06-28 16:35:11 +02:00
|
|
|
goto out;
|
2005-06-29 16:15:40 +02:00
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
rl->count[rw]++;
|
|
|
|
rl->starved[rw] = 0;
|
2005-10-28 08:29:39 +02:00
|
|
|
|
2005-10-28 08:30:39 +02:00
|
|
|
priv = !test_bit(QUEUE_FLAG_ELVSWITCH, &q->queue_flags);
|
2005-10-28 08:29:39 +02:00
|
|
|
if (priv)
|
|
|
|
rl->elvpriv++;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
|
2006-12-13 13:02:26 +01:00
|
|
|
rq = blk_alloc_request(q, rw_flags, priv, gfp_mask);
|
2005-11-12 11:09:12 +01:00
|
|
|
if (unlikely(!rq)) {
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* Allocation failed presumably due to memory. Undo anything
|
|
|
|
* we might have messed up.
|
|
|
|
*
|
|
|
|
* Allocating task should really be put onto the front of the
|
|
|
|
* wait queue, but this is pretty rare.
|
|
|
|
*/
|
|
|
|
spin_lock_irq(q->queue_lock);
|
2005-10-28 08:29:39 +02:00
|
|
|
freed_request(q, rw, priv);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* in the very unlikely event that allocation failed and no
|
|
|
|
* requests for this direction was pending, mark us starved
|
|
|
|
* so that freeing of a request in the other direction will
|
|
|
|
* notice us. another possible fix would be to split the
|
|
|
|
* rq mempool into READ and WRITE
|
|
|
|
*/
|
|
|
|
rq_starved:
|
|
|
|
if (unlikely(rl->count[rw] == 0))
|
|
|
|
rl->starved[rw] = 1;
|
|
|
|
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2005-11-12 11:09:12 +01:00
|
|
|
/*
|
|
|
|
* ioc may be NULL here, and ioc_batching will be false. That's
|
|
|
|
* OK, if the queue is under the request limit then requests need
|
|
|
|
* not count toward the nr_batch_requests limit. There will always
|
|
|
|
* be some limit enforced by BLK_BATCH_TIME.
|
|
|
|
*/
|
2005-04-17 00:20:36 +02:00
|
|
|
if (ioc_batching(q, ioc))
|
|
|
|
ioc->nr_batch_requests--;
|
2008-01-31 13:03:55 +01:00
|
|
|
|
2006-03-23 20:00:26 +01:00
|
|
|
blk_add_trace_generic(q, bio, rw, BLK_TA_GETRQ);
|
2005-04-17 00:20:36 +02:00
|
|
|
out:
|
|
|
|
return rq;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* No available requests for this queue, unplug the device and wait for some
|
|
|
|
* requests to become available.
|
2005-06-29 05:45:14 +02:00
|
|
|
*
|
|
|
|
* Called with q->queue_lock held, and returns with it unlocked.
|
2005-04-17 00:20:36 +02:00
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
static struct request *get_request_wait(struct request_queue *q, int rw_flags,
|
2005-06-27 10:55:12 +02:00
|
|
|
struct bio *bio)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2006-12-13 13:02:26 +01:00
|
|
|
const int rw = rw_flags & 0x01;
|
2005-04-17 00:20:36 +02:00
|
|
|
struct request *rq;
|
|
|
|
|
2006-12-13 13:02:26 +01:00
|
|
|
rq = get_request(q, rw_flags, bio, GFP_NOIO);
|
2005-06-29 05:45:13 +02:00
|
|
|
while (!rq) {
|
|
|
|
DEFINE_WAIT(wait);
|
2008-05-22 15:13:29 +02:00
|
|
|
struct io_context *ioc;
|
2005-04-17 00:20:36 +02:00
|
|
|
struct request_list *rl = &q->rq;
|
|
|
|
|
|
|
|
prepare_to_wait_exclusive(&rl->wait[rw], &wait,
|
|
|
|
TASK_UNINTERRUPTIBLE);
|
|
|
|
|
2008-05-22 15:13:29 +02:00
|
|
|
blk_add_trace_generic(q, bio, rw, BLK_TA_SLEEPRQ);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-05-22 15:13:29 +02:00
|
|
|
__generic_unplug_device(q);
|
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
io_schedule();
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-05-22 15:13:29 +02:00
|
|
|
/*
|
|
|
|
* After sleeping, we become a "batching" process and
|
|
|
|
* will be able to allocate at least one request, and
|
|
|
|
* up to a big batch of them for a small period time.
|
|
|
|
* See ioc_batching, ioc_set_batching
|
|
|
|
*/
|
|
|
|
ioc = current_io_context(GFP_NOIO, q->node);
|
|
|
|
ioc_set_batching(q, ioc);
|
2005-06-29 05:45:14 +02:00
|
|
|
|
2008-05-22 15:13:29 +02:00
|
|
|
spin_lock_irq(q->queue_lock);
|
2005-04-17 00:20:36 +02:00
|
|
|
finish_wait(&rl->wait[rw], &wait);
|
2008-05-22 15:13:29 +02:00
|
|
|
|
|
|
|
rq = get_request(q, rw_flags, bio, GFP_NOIO);
|
|
|
|
};
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
return rq;
|
|
|
|
}
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request *blk_get_request(struct request_queue *q, int rw, gfp_t gfp_mask)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct request *rq;
|
|
|
|
|
|
|
|
BUG_ON(rw != READ && rw != WRITE);
|
|
|
|
|
2005-06-29 05:45:14 +02:00
|
|
|
spin_lock_irq(q->queue_lock);
|
|
|
|
if (gfp_mask & __GFP_WAIT) {
|
2005-06-27 10:55:12 +02:00
|
|
|
rq = get_request_wait(q, rw, NULL);
|
2005-06-29 05:45:14 +02:00
|
|
|
} else {
|
2005-06-27 10:55:12 +02:00
|
|
|
rq = get_request(q, rw, NULL, gfp_mask);
|
2005-06-29 05:45:14 +02:00
|
|
|
if (!rq)
|
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
}
|
|
|
|
/* q->queue_lock is unlocked at this point */
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
return rq;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_get_request);
|
|
|
|
|
2006-07-20 14:54:05 +02:00
|
|
|
/**
|
|
|
|
* blk_start_queueing - initiate dispatch of requests to device
|
|
|
|
* @q: request queue to kick into gear
|
|
|
|
*
|
|
|
|
* This is basically a helper to remove the need to know whether a queue
|
|
|
|
* is plugged or not if someone just wants to initiate dispatch of requests
|
|
|
|
* for this queue.
|
|
|
|
*
|
|
|
|
* The queue lock must be held with interrupts disabled.
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
void blk_start_queueing(struct request_queue *q)
|
2006-07-20 14:54:05 +02:00
|
|
|
{
|
|
|
|
if (!blk_queue_plugged(q))
|
|
|
|
q->request_fn(q);
|
|
|
|
else
|
|
|
|
__generic_unplug_device(q);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_start_queueing);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/**
|
|
|
|
* blk_requeue_request - put a request back on queue
|
|
|
|
* @q: request queue where request should be inserted
|
|
|
|
* @rq: request to be inserted
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Drivers often keep queueing requests until the hardware cannot accept
|
|
|
|
* more, when that condition happens we need to put the request back
|
|
|
|
* on the queue. Must be called with queue lock held.
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
void blk_requeue_request(struct request_queue *q, struct request *rq)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2006-03-23 20:00:26 +01:00
|
|
|
blk_add_trace_rq(q, rq, BLK_TA_REQUEUE);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
if (blk_rq_tagged(rq))
|
|
|
|
blk_queue_end_tag(q, rq);
|
|
|
|
|
|
|
|
elv_requeue_request(q, rq);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_requeue_request);
|
|
|
|
|
|
|
|
/**
|
2008-08-19 20:13:11 +02:00
|
|
|
* blk_insert_request - insert a special request into a request queue
|
2005-04-17 00:20:36 +02:00
|
|
|
* @q: request queue where request should be inserted
|
|
|
|
* @rq: request to be inserted
|
|
|
|
* @at_head: insert request at head or tail of queue
|
|
|
|
* @data: private data
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Many block devices need to execute commands asynchronously, so they don't
|
|
|
|
* block the whole kernel from preemption during request execution. This is
|
|
|
|
* accomplished normally by inserting aritficial requests tagged as
|
2008-08-19 20:13:11 +02:00
|
|
|
* REQ_TYPE_SPECIAL in to the corresponding request queue, and letting them
|
|
|
|
* be scheduled for actual execution by the request queue.
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
|
|
|
* We have the option of inserting the head or the tail of the queue.
|
|
|
|
* Typically we use the tail for new ioctls and so forth. We use the head
|
|
|
|
* of the queue for things like a QUEUE_FULL message from a device, or a
|
|
|
|
* host that is unable to accept a particular command.
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
void blk_insert_request(struct request_queue *q, struct request *rq,
|
2005-04-24 09:06:05 +02:00
|
|
|
int at_head, void *data)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2005-04-24 09:06:05 +02:00
|
|
|
int where = at_head ? ELEVATOR_INSERT_FRONT : ELEVATOR_INSERT_BACK;
|
2005-04-17 00:20:36 +02:00
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* tell I/O scheduler that this isn't a regular read/write (ie it
|
|
|
|
* must not attempt merges on this) and that it acts as a soft
|
|
|
|
* barrier
|
|
|
|
*/
|
2006-08-10 08:44:47 +02:00
|
|
|
rq->cmd_type = REQ_TYPE_SPECIAL;
|
|
|
|
rq->cmd_flags |= REQ_SOFTBARRIER;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
rq->special = data;
|
|
|
|
|
|
|
|
spin_lock_irqsave(q->queue_lock, flags);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If command is tagged, release the tag
|
|
|
|
*/
|
2005-04-24 09:06:05 +02:00
|
|
|
if (blk_rq_tagged(rq))
|
|
|
|
blk_queue_end_tag(q, rq);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-10-23 15:05:46 +02:00
|
|
|
drive_stat_acct(rq, 1);
|
2005-04-24 09:06:05 +02:00
|
|
|
__elv_add_request(q, rq, where, 0);
|
2006-07-20 14:54:05 +02:00
|
|
|
blk_start_queueing(q);
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_unlock_irqrestore(q->queue_lock, flags);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_insert_request);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* add-request adds a request to the linked list.
|
|
|
|
* queue lock is held and interrupts disabled, as we muck with the
|
|
|
|
* request queue list.
|
|
|
|
*/
|
2008-01-31 13:03:55 +01:00
|
|
|
static inline void add_request(struct request_queue *q, struct request *req)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-10-23 15:05:46 +02:00
|
|
|
drive_stat_acct(req, 1);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* elevator indicated where it wants this request to be
|
|
|
|
* inserted at elevator_merge time
|
|
|
|
*/
|
|
|
|
__elv_add_request(q, req, ELEVATOR_INSERT_SORT, 0);
|
|
|
|
}
|
2008-01-31 13:03:55 +01:00
|
|
|
|
2008-08-25 12:56:14 +02:00
|
|
|
static void part_round_stats_single(int cpu, struct hd_struct *part,
|
|
|
|
unsigned long now)
|
|
|
|
{
|
|
|
|
if (now == part->stamp)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (part->in_flight) {
|
|
|
|
__part_stat_add(cpu, part, time_in_queue,
|
|
|
|
part->in_flight * (now - part->stamp));
|
|
|
|
__part_stat_add(cpu, part, io_ticks, (now - part->stamp));
|
|
|
|
}
|
|
|
|
part->stamp = now;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* part_round_stats() - Round off the performance stats on a struct
|
2005-04-17 00:20:36 +02:00
|
|
|
* disk_stats.
|
|
|
|
*
|
|
|
|
* The average IO queue length and utilisation statistics are maintained
|
|
|
|
* by observing the current state of the queue length and the amount of
|
|
|
|
* time it has been in this state for.
|
|
|
|
*
|
|
|
|
* Normally, that accounting is done on IO completion, but that can result
|
|
|
|
* in more than a second's worth of IO being accounted for within any one
|
|
|
|
* second, leading to >100% utilisation. To deal with that, we call this
|
|
|
|
* function to do a round-off before returning the results when reading
|
|
|
|
* /proc/diskstats. This accounts immediately for all queue usage up to
|
|
|
|
* the current jiffies and restarts the counters again.
|
|
|
|
*/
|
2008-08-25 12:47:21 +02:00
|
|
|
void part_round_stats(int cpu, struct hd_struct *part)
|
2008-02-08 11:04:35 +01:00
|
|
|
{
|
|
|
|
unsigned long now = jiffies;
|
|
|
|
|
2008-08-25 12:56:14 +02:00
|
|
|
if (part->partno)
|
|
|
|
part_round_stats_single(cpu, &part_to_disk(part)->part0, now);
|
|
|
|
part_round_stats_single(cpu, part, now);
|
2008-02-08 11:04:35 +01:00
|
|
|
}
|
2008-08-25 12:56:14 +02:00
|
|
|
EXPORT_SYMBOL_GPL(part_round_stats);
|
2008-02-08 11:04:35 +01:00
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* queue lock must be held
|
|
|
|
*/
|
2007-07-24 09:28:11 +02:00
|
|
|
void __blk_put_request(struct request_queue *q, struct request *req)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
if (unlikely(!q))
|
|
|
|
return;
|
|
|
|
if (unlikely(--req->ref_count))
|
|
|
|
return;
|
|
|
|
|
2005-10-20 16:23:44 +02:00
|
|
|
elv_completed_request(q, req);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* Request may not have originated from ll_rw_blk. if not,
|
|
|
|
* it didn't come out of our reserved rq pools
|
|
|
|
*/
|
2006-08-10 08:59:11 +02:00
|
|
|
if (req->cmd_flags & REQ_ALLOCED) {
|
2005-04-17 00:20:36 +02:00
|
|
|
int rw = rq_data_dir(req);
|
2006-08-10 08:44:47 +02:00
|
|
|
int priv = req->cmd_flags & REQ_ELVPRIV;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
BUG_ON(!list_empty(&req->queuelist));
|
2006-07-28 09:23:08 +02:00
|
|
|
BUG_ON(!hlist_unhashed(&req->hash));
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
blk_free_request(q, req);
|
2005-10-28 08:29:39 +02:00
|
|
|
freed_request(q, rw, priv);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
}
|
2005-11-11 12:30:24 +01:00
|
|
|
EXPORT_SYMBOL_GPL(__blk_put_request);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
void blk_put_request(struct request *req)
|
|
|
|
{
|
2005-10-20 16:23:44 +02:00
|
|
|
unsigned long flags;
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q = req->q;
|
2005-10-20 16:23:44 +02:00
|
|
|
|
2008-07-15 21:21:45 +02:00
|
|
|
spin_lock_irqsave(q->queue_lock, flags);
|
|
|
|
__blk_put_request(q, req);
|
|
|
|
spin_unlock_irqrestore(q->queue_lock, flags);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_put_request);
|
|
|
|
|
2008-01-29 14:53:40 +01:00
|
|
|
void init_request_from_bio(struct request *req, struct bio *bio)
|
2006-01-06 09:49:58 +01:00
|
|
|
{
|
2008-09-13 20:26:01 +02:00
|
|
|
req->cpu = bio->bi_comp_cpu;
|
2006-08-10 08:44:47 +02:00
|
|
|
req->cmd_type = REQ_TYPE_FS;
|
2006-01-06 09:49:58 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
* inherit FAILFAST from bio (for read-ahead, and explicit FAILFAST)
|
|
|
|
*/
|
|
|
|
if (bio_rw_ahead(bio) || bio_failfast(bio))
|
2006-08-10 08:44:47 +02:00
|
|
|
req->cmd_flags |= REQ_FAILFAST;
|
2006-01-06 09:49:58 +01:00
|
|
|
|
|
|
|
/*
|
|
|
|
* REQ_BARRIER implies no merging, but lets make it explicit
|
|
|
|
*/
|
2008-08-05 19:01:53 +02:00
|
|
|
if (unlikely(bio_discard(bio))) {
|
2008-08-09 17:42:20 +02:00
|
|
|
req->cmd_flags |= REQ_DISCARD;
|
|
|
|
if (bio_barrier(bio))
|
|
|
|
req->cmd_flags |= REQ_SOFTBARRIER;
|
2008-08-05 19:01:53 +02:00
|
|
|
req->q->prepare_discard_fn(req->q, req);
|
2008-08-09 17:42:20 +02:00
|
|
|
} else if (unlikely(bio_barrier(bio)))
|
|
|
|
req->cmd_flags |= (REQ_HARDBARRIER | REQ_NOMERGE);
|
2006-01-06 09:49:58 +01:00
|
|
|
|
2006-06-13 08:26:10 +02:00
|
|
|
if (bio_sync(bio))
|
2006-08-10 08:44:47 +02:00
|
|
|
req->cmd_flags |= REQ_RW_SYNC;
|
2006-08-10 09:01:02 +02:00
|
|
|
if (bio_rw_meta(bio))
|
|
|
|
req->cmd_flags |= REQ_RW_META;
|
2006-06-13 08:26:10 +02:00
|
|
|
|
2006-01-06 09:49:58 +01:00
|
|
|
req->errors = 0;
|
|
|
|
req->hard_sector = req->sector = bio->bi_sector;
|
|
|
|
req->ioprio = bio_prio(bio);
|
|
|
|
req->start_time = jiffies;
|
2007-08-16 13:31:30 +02:00
|
|
|
blk_rq_bio_prep(req->q, req, bio);
|
2006-01-06 09:49:58 +01:00
|
|
|
}
|
|
|
|
|
2007-07-24 09:28:11 +02:00
|
|
|
static int __make_request(struct request_queue *q, struct bio *bio)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2005-06-29 05:45:13 +02:00
|
|
|
struct request *req;
|
2008-08-05 19:01:53 +02:00
|
|
|
int el_ret, nr_sectors, barrier, discard, err;
|
2006-07-18 04:14:45 +02:00
|
|
|
const unsigned short prio = bio_prio(bio);
|
|
|
|
const int sync = bio_sync(bio);
|
2006-12-13 13:02:26 +01:00
|
|
|
int rw_flags;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
nr_sectors = bio_sectors(bio);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* low level driver can indicate that it wants pages above a
|
|
|
|
* certain limit bounced to low memory (ie for highmem, or even
|
|
|
|
* ISA dma in theory)
|
|
|
|
*/
|
|
|
|
blk_queue_bounce(q, &bio);
|
|
|
|
|
|
|
|
barrier = bio_barrier(bio);
|
2008-08-09 17:42:20 +02:00
|
|
|
if (unlikely(barrier) && bio_has_data(bio) &&
|
|
|
|
(q->next_ordered == QUEUE_ORDERED_NONE)) {
|
2005-04-17 00:20:36 +02:00
|
|
|
err = -EOPNOTSUPP;
|
|
|
|
goto end_io;
|
|
|
|
}
|
|
|
|
|
2008-08-05 19:01:53 +02:00
|
|
|
discard = bio_discard(bio);
|
|
|
|
if (unlikely(discard) && !q->prepare_discard_fn) {
|
|
|
|
err = -EOPNOTSUPP;
|
|
|
|
goto end_io;
|
|
|
|
}
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
spin_lock_irq(q->queue_lock);
|
|
|
|
|
2005-06-29 05:45:13 +02:00
|
|
|
if (unlikely(barrier) || elv_queue_empty(q))
|
2005-04-17 00:20:36 +02:00
|
|
|
goto get_rq;
|
|
|
|
|
|
|
|
el_ret = elv_merge(q, &req, bio);
|
|
|
|
switch (el_ret) {
|
2008-01-31 13:03:55 +01:00
|
|
|
case ELEVATOR_BACK_MERGE:
|
|
|
|
BUG_ON(!rq_mergeable(req));
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
if (!ll_back_merge_fn(q, req, bio))
|
|
|
|
break;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
blk_add_trace_bio(q, bio, BLK_TA_BACKMERGE);
|
2006-03-23 20:00:26 +01:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
req->biotail->bi_next = bio;
|
|
|
|
req->biotail = bio;
|
|
|
|
req->nr_sectors = req->hard_nr_sectors += nr_sectors;
|
|
|
|
req->ioprio = ioprio_best(req->ioprio, prio);
|
2008-08-26 10:25:02 +02:00
|
|
|
if (!blk_rq_cpu_valid(req))
|
|
|
|
req->cpu = bio->bi_comp_cpu;
|
2008-01-31 13:03:55 +01:00
|
|
|
drive_stat_acct(req, 0);
|
|
|
|
if (!attempt_back_merge(q, req))
|
|
|
|
elv_merged_request(q, req, el_ret);
|
|
|
|
goto out;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
case ELEVATOR_FRONT_MERGE:
|
|
|
|
BUG_ON(!rq_mergeable(req));
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
if (!ll_front_merge_fn(q, req, bio))
|
|
|
|
break;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
blk_add_trace_bio(q, bio, BLK_TA_FRONTMERGE);
|
2006-03-23 20:00:26 +01:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
bio->bi_next = req->bio;
|
|
|
|
req->bio = bio;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
/*
|
|
|
|
* may not be valid. if the low level driver said
|
|
|
|
* it didn't need a bounce buffer then it better
|
|
|
|
* not touch req->buffer either...
|
|
|
|
*/
|
|
|
|
req->buffer = bio_data(bio);
|
|
|
|
req->current_nr_sectors = bio_cur_sectors(bio);
|
|
|
|
req->hard_cur_sectors = req->current_nr_sectors;
|
|
|
|
req->sector = req->hard_sector = bio->bi_sector;
|
|
|
|
req->nr_sectors = req->hard_nr_sectors += nr_sectors;
|
|
|
|
req->ioprio = ioprio_best(req->ioprio, prio);
|
2008-08-26 10:25:02 +02:00
|
|
|
if (!blk_rq_cpu_valid(req))
|
|
|
|
req->cpu = bio->bi_comp_cpu;
|
2008-01-31 13:03:55 +01:00
|
|
|
drive_stat_acct(req, 0);
|
|
|
|
if (!attempt_front_merge(q, req))
|
|
|
|
elv_merged_request(q, req, el_ret);
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
/* ELV_NO_MERGE: elevator says don't/can't merge. */
|
|
|
|
default:
|
|
|
|
;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2005-06-29 05:45:13 +02:00
|
|
|
get_rq:
|
2006-12-13 13:02:26 +01:00
|
|
|
/*
|
|
|
|
* This sync check and mask will be re-done in init_request_from_bio(),
|
|
|
|
* but we need to set it earlier to expose the sync flag to the
|
|
|
|
* rq allocator and io schedulers.
|
|
|
|
*/
|
|
|
|
rw_flags = bio_data_dir(bio);
|
|
|
|
if (sync)
|
|
|
|
rw_flags |= REQ_RW_SYNC;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
2005-06-29 05:45:13 +02:00
|
|
|
* Grab a free request. This is might sleep but can not fail.
|
2005-06-29 05:45:14 +02:00
|
|
|
* Returns with the queue unlocked.
|
2005-06-29 05:45:13 +02:00
|
|
|
*/
|
2006-12-13 13:02:26 +01:00
|
|
|
req = get_request_wait(q, rw_flags, bio);
|
2005-06-29 05:45:14 +02:00
|
|
|
|
2005-06-29 05:45:13 +02:00
|
|
|
/*
|
|
|
|
* After dropping the lock and possibly sleeping here, our request
|
|
|
|
* may now be mergeable after it had proven unmergeable (above).
|
|
|
|
* We don't worry about that case for efficiency. It won't happen
|
|
|
|
* often, and the elevators are able to handle it.
|
2005-04-17 00:20:36 +02:00
|
|
|
*/
|
2006-01-06 09:49:58 +01:00
|
|
|
init_request_from_bio(req, bio);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2005-06-29 05:45:13 +02:00
|
|
|
spin_lock_irq(q->queue_lock);
|
2008-09-13 20:26:01 +02:00
|
|
|
if (test_bit(QUEUE_FLAG_SAME_COMP, &q->queue_flags) ||
|
|
|
|
bio_flagged(bio, BIO_CPU_AFFINE))
|
|
|
|
req->cpu = blk_cpu_to_group(smp_processor_id());
|
2005-06-29 05:45:13 +02:00
|
|
|
if (elv_queue_empty(q))
|
|
|
|
blk_plug_device(q);
|
2005-04-17 00:20:36 +02:00
|
|
|
add_request(q, req);
|
|
|
|
out:
|
2005-04-17 00:25:40 +02:00
|
|
|
if (sync)
|
2005-04-17 00:20:36 +02:00
|
|
|
__generic_unplug_device(q);
|
|
|
|
spin_unlock_irq(q->queue_lock);
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
end_io:
|
2007-09-27 12:47:43 +02:00
|
|
|
bio_endio(bio, err);
|
2005-04-17 00:20:36 +02:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If bio->bi_dev is a partition, remap the location
|
|
|
|
*/
|
|
|
|
static inline void blk_partition_remap(struct bio *bio)
|
|
|
|
{
|
|
|
|
struct block_device *bdev = bio->bi_bdev;
|
|
|
|
|
2007-09-27 13:01:25 +02:00
|
|
|
if (bio_sectors(bio) && bdev != bdev->bd_contains) {
|
2005-04-17 00:20:36 +02:00
|
|
|
struct hd_struct *p = bdev->bd_part;
|
|
|
|
|
|
|
|
bio->bi_sector += p->start_sect;
|
|
|
|
bio->bi_bdev = bdev->bd_contains;
|
2007-08-07 15:30:23 +02:00
|
|
|
|
|
|
|
blk_add_trace_remap(bdev_get_queue(bio->bi_bdev), bio,
|
|
|
|
bdev->bd_dev, bio->bi_sector,
|
|
|
|
bio->bi_sector - p->start_sect);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void handle_bad_sector(struct bio *bio)
|
|
|
|
{
|
|
|
|
char b[BDEVNAME_SIZE];
|
|
|
|
|
|
|
|
printk(KERN_INFO "attempt to access beyond end of device\n");
|
|
|
|
printk(KERN_INFO "%s: rw=%ld, want=%Lu, limit=%Lu\n",
|
|
|
|
bdevname(bio->bi_bdev, b),
|
|
|
|
bio->bi_rw,
|
|
|
|
(unsigned long long)bio->bi_sector + bio_sectors(bio),
|
|
|
|
(long long)(bio->bi_bdev->bd_inode->i_size >> 9));
|
|
|
|
|
|
|
|
set_bit(BIO_EOF, &bio->bi_flags);
|
|
|
|
}
|
|
|
|
|
2006-12-08 11:39:46 +01:00
|
|
|
#ifdef CONFIG_FAIL_MAKE_REQUEST
|
|
|
|
|
|
|
|
static DECLARE_FAULT_ATTR(fail_make_request);
|
|
|
|
|
|
|
|
static int __init setup_fail_make_request(char *str)
|
|
|
|
{
|
|
|
|
return setup_fault_attr(&fail_make_request, str);
|
|
|
|
}
|
|
|
|
__setup("fail_make_request=", setup_fail_make_request);
|
|
|
|
|
|
|
|
static int should_fail_request(struct bio *bio)
|
|
|
|
{
|
2008-08-25 12:56:13 +02:00
|
|
|
struct hd_struct *part = bio->bi_bdev->bd_part;
|
|
|
|
|
|
|
|
if (part_to_disk(part)->part0.make_it_fail || part->make_it_fail)
|
2006-12-08 11:39:46 +01:00
|
|
|
return should_fail(&fail_make_request, bio->bi_size);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __init fail_make_request_debugfs(void)
|
|
|
|
{
|
|
|
|
return init_fault_attr_dentries(&fail_make_request,
|
|
|
|
"fail_make_request");
|
|
|
|
}
|
|
|
|
|
|
|
|
late_initcall(fail_make_request_debugfs);
|
|
|
|
|
|
|
|
#else /* CONFIG_FAIL_MAKE_REQUEST */
|
|
|
|
|
|
|
|
static inline int should_fail_request(struct bio *bio)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif /* CONFIG_FAIL_MAKE_REQUEST */
|
|
|
|
|
2007-07-18 13:27:58 +02:00
|
|
|
/*
|
|
|
|
* Check whether this bio extends beyond the end of the device.
|
|
|
|
*/
|
|
|
|
static inline int bio_check_eod(struct bio *bio, unsigned int nr_sectors)
|
|
|
|
{
|
|
|
|
sector_t maxsector;
|
|
|
|
|
|
|
|
if (!nr_sectors)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* Test device or partition size, when known. */
|
|
|
|
maxsector = bio->bi_bdev->bd_inode->i_size >> 9;
|
|
|
|
if (maxsector) {
|
|
|
|
sector_t sector = bio->bi_sector;
|
|
|
|
|
|
|
|
if (maxsector < nr_sectors || maxsector - nr_sectors < sector) {
|
|
|
|
/*
|
|
|
|
* This may well happen - the kernel calls bread()
|
|
|
|
* without checking the size of the device, e.g., when
|
|
|
|
* mounting a device.
|
|
|
|
*/
|
|
|
|
handle_bad_sector(bio);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/**
|
2008-08-19 20:13:11 +02:00
|
|
|
* generic_make_request - hand a buffer to its device driver for I/O
|
2005-04-17 00:20:36 +02:00
|
|
|
* @bio: The bio describing the location in memory and on the device.
|
|
|
|
*
|
|
|
|
* generic_make_request() is used to make I/O requests of block
|
|
|
|
* devices. It is passed a &struct bio, which describes the I/O that needs
|
|
|
|
* to be done.
|
|
|
|
*
|
|
|
|
* generic_make_request() does not return any status. The
|
|
|
|
* success/failure status of the request, along with notification of
|
|
|
|
* completion, is delivered asynchronously through the bio->bi_end_io
|
|
|
|
* function described (one day) else where.
|
|
|
|
*
|
|
|
|
* The caller of generic_make_request must make sure that bi_io_vec
|
|
|
|
* are set to describe the memory buffer, and that bi_dev and bi_sector are
|
|
|
|
* set to describe the device address, and the
|
|
|
|
* bi_end_io and optionally bi_private are set to describe how
|
|
|
|
* completion notification should be signaled.
|
|
|
|
*
|
|
|
|
* generic_make_request and the drivers it calls may use bi_next if this
|
|
|
|
* bio happens to be merged with someone else, and may change bi_dev and
|
|
|
|
* bi_sector for remaps as it sees fit. So the values of these fields
|
|
|
|
* should NOT be depended on after the call to generic_make_request.
|
|
|
|
*/
|
When stacked block devices are in-use (e.g. md or dm), the recursive calls
to generic_make_request can use up a lot of space, and we would rather they
didn't.
As generic_make_request is a void function, and as it is generally not
expected that it will have any effect immediately, it is safe to delay any
call to generic_make_request until there is sufficient stack space
available.
As ->bi_next is reserved for the driver to use, it can have no valid value
when generic_make_request is called, and as __make_request implicitly
assumes it will be NULL (ELEVATOR_BACK_MERGE fork of switch) we can be
certain that all callers set it to NULL. We can therefore safely use
bi_next to link pending requests together, providing we clear it before
making the real call.
So, we choose to allow each thread to only be active in one
generic_make_request at a time. If a subsequent (recursive) call is made,
the bio is linked into a per-thread list, and is handled when the active
call completes.
As the list of pending bios is per-thread, there are no locking issues to
worry about.
I say above that it is "safe to delay any call...". There are, however,
some behaviours of a make_request_fn which would make it unsafe. These
include any behaviour that assumes anything will have changed after a
recursive call to generic_make_request.
These could include:
- waiting for that call to finish and call it's bi_end_io function.
md use to sometimes do this (marking the superblock dirty before
completing a write) but doesn't any more
- inspecting the bio for fields that generic_make_request might
change, such as bi_sector or bi_bdev. It is hard to see a good
reason for this, and I don't think anyone actually does it.
- inspecing the queue to see if, e.g. it is 'full' yet. Again, I
think this is very unlikely to be useful, or to be done.
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: <dm-devel@redhat.com>
Alasdair G Kergon <agk@redhat.com> said:
I can see nothing wrong with this in principle.
For device-mapper at the moment though it's essential that, while the bio
mappings may now get delayed, they still get processed in exactly
the same order as they were passed to generic_make_request().
My main concern is whether the timing changes implicit in this patch
will make the rare data-corrupting races in the existing snapshot code
more likely. (I'm working on a fix for these races, but the unfinished
patch is already several hundred lines long.)
It would be helpful if some people on this mailing list would test
this patch in various scenarios and report back.
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2007-05-01 09:53:42 +02:00
|
|
|
static inline void __generic_make_request(struct bio *bio)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-07-24 09:28:11 +02:00
|
|
|
struct request_queue *q;
|
2006-10-31 07:07:21 +01:00
|
|
|
sector_t old_sector;
|
2005-04-17 00:20:36 +02:00
|
|
|
int ret, nr_sectors = bio_sectors(bio);
|
2006-03-23 20:00:26 +01:00
|
|
|
dev_t old_dev;
|
2007-11-02 08:49:08 +01:00
|
|
|
int err = -EIO;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
might_sleep();
|
|
|
|
|
2007-07-18 13:27:58 +02:00
|
|
|
if (bio_check_eod(bio, nr_sectors))
|
|
|
|
goto end_io;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Resolve the mapping until finished. (drivers are
|
|
|
|
* still free to implement/resolve their own stacking
|
|
|
|
* by explicitly returning 0)
|
|
|
|
*
|
|
|
|
* NOTE: we don't repeat the blk_size check for each new device.
|
|
|
|
* Stacking drivers are expected to know what they are doing.
|
|
|
|
*/
|
2006-10-31 07:07:21 +01:00
|
|
|
old_sector = -1;
|
2006-03-23 20:00:26 +01:00
|
|
|
old_dev = 0;
|
2005-04-17 00:20:36 +02:00
|
|
|
do {
|
|
|
|
char b[BDEVNAME_SIZE];
|
|
|
|
|
|
|
|
q = bdev_get_queue(bio->bi_bdev);
|
|
|
|
if (!q) {
|
|
|
|
printk(KERN_ERR
|
|
|
|
"generic_make_request: Trying to access "
|
|
|
|
"nonexistent block-device %s (%Lu)\n",
|
|
|
|
bdevname(bio->bi_bdev, b),
|
|
|
|
(long long) bio->bi_sector);
|
|
|
|
end_io:
|
2007-11-02 08:49:08 +01:00
|
|
|
bio_endio(bio, err);
|
2005-04-17 00:20:36 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
2007-07-18 13:13:10 +02:00
|
|
|
if (unlikely(nr_sectors > q->max_hw_sectors)) {
|
2008-01-31 13:03:55 +01:00
|
|
|
printk(KERN_ERR "bio too big device %s (%u > %u)\n",
|
2005-04-17 00:20:36 +02:00
|
|
|
bdevname(bio->bi_bdev, b),
|
|
|
|
bio_sectors(bio),
|
|
|
|
q->max_hw_sectors);
|
|
|
|
goto end_io;
|
|
|
|
}
|
|
|
|
|
2005-06-23 09:08:53 +02:00
|
|
|
if (unlikely(test_bit(QUEUE_FLAG_DEAD, &q->queue_flags)))
|
2005-04-17 00:20:36 +02:00
|
|
|
goto end_io;
|
|
|
|
|
2006-12-08 11:39:46 +01:00
|
|
|
if (should_fail_request(bio))
|
|
|
|
goto end_io;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* If this device has partitions, remap block n
|
|
|
|
* of partition p to block n+start(p) of the disk.
|
|
|
|
*/
|
|
|
|
blk_partition_remap(bio);
|
|
|
|
|
2008-06-30 20:04:41 +02:00
|
|
|
if (bio_integrity_enabled(bio) && bio_integrity_prep(bio))
|
|
|
|
goto end_io;
|
|
|
|
|
2006-10-31 07:07:21 +01:00
|
|
|
if (old_sector != -1)
|
2007-07-18 13:13:10 +02:00
|
|
|
blk_add_trace_remap(q, bio, old_dev, bio->bi_sector,
|
2006-10-31 07:07:21 +01:00
|
|
|
old_sector);
|
2006-03-23 20:00:26 +01:00
|
|
|
|
|
|
|
blk_add_trace_bio(q, bio, BLK_TA_QUEUE);
|
|
|
|
|
2006-10-31 07:07:21 +01:00
|
|
|
old_sector = bio->bi_sector;
|
2006-03-23 20:00:26 +01:00
|
|
|
old_dev = bio->bi_bdev->bd_dev;
|
|
|
|
|
2007-07-18 13:27:58 +02:00
|
|
|
if (bio_check_eod(bio, nr_sectors))
|
|
|
|
goto end_io;
|
2008-08-05 19:01:53 +02:00
|
|
|
if ((bio_empty_barrier(bio) && !q->prepare_flush_fn) ||
|
|
|
|
(bio_discard(bio) && !q->prepare_discard_fn)) {
|
2007-11-02 08:49:08 +01:00
|
|
|
err = -EOPNOTSUPP;
|
|
|
|
goto end_io;
|
|
|
|
}
|
2006-10-31 07:07:21 +01:00
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
ret = q->make_request_fn(q, bio);
|
|
|
|
} while (ret);
|
|
|
|
}
|
|
|
|
|
When stacked block devices are in-use (e.g. md or dm), the recursive calls
to generic_make_request can use up a lot of space, and we would rather they
didn't.
As generic_make_request is a void function, and as it is generally not
expected that it will have any effect immediately, it is safe to delay any
call to generic_make_request until there is sufficient stack space
available.
As ->bi_next is reserved for the driver to use, it can have no valid value
when generic_make_request is called, and as __make_request implicitly
assumes it will be NULL (ELEVATOR_BACK_MERGE fork of switch) we can be
certain that all callers set it to NULL. We can therefore safely use
bi_next to link pending requests together, providing we clear it before
making the real call.
So, we choose to allow each thread to only be active in one
generic_make_request at a time. If a subsequent (recursive) call is made,
the bio is linked into a per-thread list, and is handled when the active
call completes.
As the list of pending bios is per-thread, there are no locking issues to
worry about.
I say above that it is "safe to delay any call...". There are, however,
some behaviours of a make_request_fn which would make it unsafe. These
include any behaviour that assumes anything will have changed after a
recursive call to generic_make_request.
These could include:
- waiting for that call to finish and call it's bi_end_io function.
md use to sometimes do this (marking the superblock dirty before
completing a write) but doesn't any more
- inspecting the bio for fields that generic_make_request might
change, such as bi_sector or bi_bdev. It is hard to see a good
reason for this, and I don't think anyone actually does it.
- inspecing the queue to see if, e.g. it is 'full' yet. Again, I
think this is very unlikely to be useful, or to be done.
Signed-off-by: Neil Brown <neilb@suse.de>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: <dm-devel@redhat.com>
Alasdair G Kergon <agk@redhat.com> said:
I can see nothing wrong with this in principle.
For device-mapper at the moment though it's essential that, while the bio
mappings may now get delayed, they still get processed in exactly
the same order as they were passed to generic_make_request().
My main concern is whether the timing changes implicit in this patch
will make the rare data-corrupting races in the existing snapshot code
more likely. (I'm working on a fix for these races, but the unfinished
patch is already several hundred lines long.)
It would be helpful if some people on this mailing list would test
this patch in various scenarios and report back.
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2007-05-01 09:53:42 +02:00
|
|
|
/*
|
|
|
|
* We only want one ->make_request_fn to be active at a time,
|
|
|
|
* else stack usage with stacked devices could be a problem.
|
|
|
|
* So use current->bio_{list,tail} to keep a list of requests
|
|
|
|
* submited by a make_request_fn function.
|
|
|
|
* current->bio_tail is also used as a flag to say if
|
|
|
|
* generic_make_request is currently active in this task or not.
|
|
|
|
* If it is NULL, then no make_request is active. If it is non-NULL,
|
|
|
|
* then a make_request is active, and new requests should be added
|
|
|
|
* at the tail
|
|
|
|
*/
|
|
|
|
void generic_make_request(struct bio *bio)
|
|
|
|
{
|
|
|
|
if (current->bio_tail) {
|
|
|
|
/* make_request is active */
|
|
|
|
*(current->bio_tail) = bio;
|
|
|
|
bio->bi_next = NULL;
|
|
|
|
current->bio_tail = &bio->bi_next;
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
/* following loop may be a bit non-obvious, and so deserves some
|
|
|
|
* explanation.
|
|
|
|
* Before entering the loop, bio->bi_next is NULL (as all callers
|
|
|
|
* ensure that) so we have a list with a single bio.
|
|
|
|
* We pretend that we have just taken it off a longer list, so
|
|
|
|
* we assign bio_list to the next (which is NULL) and bio_tail
|
|
|
|
* to &bio_list, thus initialising the bio_list of new bios to be
|
|
|
|
* added. __generic_make_request may indeed add some more bios
|
|
|
|
* through a recursive call to generic_make_request. If it
|
|
|
|
* did, we find a non-NULL value in bio_list and re-enter the loop
|
|
|
|
* from the top. In this case we really did just take the bio
|
|
|
|
* of the top of the list (no pretending) and so fixup bio_list and
|
|
|
|
* bio_tail or bi_next, and call into __generic_make_request again.
|
|
|
|
*
|
|
|
|
* The loop was structured like this to make only one call to
|
|
|
|
* __generic_make_request (which is important as it is large and
|
|
|
|
* inlined) and to keep the structure simple.
|
|
|
|
*/
|
|
|
|
BUG_ON(bio->bi_next);
|
|
|
|
do {
|
|
|
|
current->bio_list = bio->bi_next;
|
|
|
|
if (bio->bi_next == NULL)
|
|
|
|
current->bio_tail = ¤t->bio_list;
|
|
|
|
else
|
|
|
|
bio->bi_next = NULL;
|
|
|
|
__generic_make_request(bio);
|
|
|
|
bio = current->bio_list;
|
|
|
|
} while (bio);
|
|
|
|
current->bio_tail = NULL; /* deactivate */
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
EXPORT_SYMBOL(generic_make_request);
|
|
|
|
|
|
|
|
/**
|
2008-08-19 20:13:11 +02:00
|
|
|
* submit_bio - submit a bio to the block device layer for I/O
|
2005-04-17 00:20:36 +02:00
|
|
|
* @rw: whether to %READ or %WRITE, or maybe to %READA (read ahead)
|
|
|
|
* @bio: The &struct bio which describes the I/O
|
|
|
|
*
|
|
|
|
* submit_bio() is very similar in purpose to generic_make_request(), and
|
|
|
|
* uses that function to do most of the work. Both are fairly rough
|
2008-08-19 20:13:11 +02:00
|
|
|
* interfaces; @bio must be presetup and ready for I/O.
|
2005-04-17 00:20:36 +02:00
|
|
|
*
|
|
|
|
*/
|
|
|
|
void submit_bio(int rw, struct bio *bio)
|
|
|
|
{
|
|
|
|
int count = bio_sectors(bio);
|
|
|
|
|
2005-06-27 10:55:12 +02:00
|
|
|
bio->bi_rw |= rw;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2007-09-27 13:01:25 +02:00
|
|
|
/*
|
|
|
|
* If it's a regular read/write or a barrier with data attached,
|
|
|
|
* go through the normal accounting stuff before submission.
|
|
|
|
*/
|
2008-08-08 11:04:44 +02:00
|
|
|
if (bio_has_data(bio)) {
|
2007-09-27 13:01:25 +02:00
|
|
|
if (rw & WRITE) {
|
|
|
|
count_vm_events(PGPGOUT, count);
|
|
|
|
} else {
|
|
|
|
task_io_account_read(bio->bi_size);
|
|
|
|
count_vm_events(PGPGIN, count);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (unlikely(block_dump)) {
|
|
|
|
char b[BDEVNAME_SIZE];
|
|
|
|
printk(KERN_DEBUG "%s(%d): %s block %Lu on %s\n",
|
2007-10-19 08:40:40 +02:00
|
|
|
current->comm, task_pid_nr(current),
|
2007-09-27 13:01:25 +02:00
|
|
|
(rw & WRITE) ? "WRITE" : "READ",
|
|
|
|
(unsigned long long)bio->bi_sector,
|
2008-01-31 13:03:55 +01:00
|
|
|
bdevname(bio->bi_bdev, b));
|
2007-09-27 13:01:25 +02:00
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
generic_make_request(bio);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(submit_bio);
|
|
|
|
|
2007-12-11 23:52:28 +01:00
|
|
|
/**
|
|
|
|
* __end_that_request_first - end I/O on a request
|
|
|
|
* @req: the request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @error: %0 for success, < %0 for error
|
2007-12-11 23:52:28 +01:00
|
|
|
* @nr_bytes: number of bytes to complete
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Ends I/O on a number of bytes attached to @req, and sets it up
|
|
|
|
* for the next range of segments (if any) in the cluster.
|
|
|
|
*
|
|
|
|
* Return:
|
2008-08-19 20:13:11 +02:00
|
|
|
* %0 - we are done with this request, call end_that_request_last()
|
|
|
|
* %1 - still buffers pending for this request
|
2007-12-11 23:52:28 +01:00
|
|
|
**/
|
2007-12-11 23:53:03 +01:00
|
|
|
static int __end_that_request_first(struct request *req, int error,
|
2005-04-17 00:20:36 +02:00
|
|
|
int nr_bytes)
|
|
|
|
{
|
2007-12-11 23:53:03 +01:00
|
|
|
int total_bytes, bio_nbytes, next_idx = 0;
|
2005-04-17 00:20:36 +02:00
|
|
|
struct bio *bio;
|
|
|
|
|
2006-03-23 20:00:26 +01:00
|
|
|
blk_add_trace_rq(req->q, req, BLK_TA_COMPLETE);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
2008-08-19 20:13:11 +02:00
|
|
|
* for a REQ_TYPE_BLOCK_PC request, we want to carry any eventual
|
2005-04-17 00:20:36 +02:00
|
|
|
* sense key with us all the way through
|
|
|
|
*/
|
|
|
|
if (!blk_pc_request(req))
|
|
|
|
req->errors = 0;
|
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
if (error && (blk_fs_request(req) && !(req->cmd_flags & REQ_QUIET))) {
|
|
|
|
printk(KERN_ERR "end_request: I/O error, dev %s, sector %llu\n",
|
2005-04-17 00:20:36 +02:00
|
|
|
req->rq_disk ? req->rq_disk->disk_name : "?",
|
|
|
|
(unsigned long long)req->sector);
|
|
|
|
}
|
|
|
|
|
2005-11-01 08:35:42 +01:00
|
|
|
if (blk_fs_request(req) && req->rq_disk) {
|
2005-11-01 09:26:16 +01:00
|
|
|
const int rw = rq_data_dir(req);
|
2008-09-03 09:03:02 +02:00
|
|
|
struct hd_struct *part;
|
2008-08-25 12:47:21 +02:00
|
|
|
int cpu;
|
2005-11-01 09:26:16 +01:00
|
|
|
|
2008-08-25 12:56:14 +02:00
|
|
|
cpu = part_stat_lock();
|
2008-09-03 09:03:02 +02:00
|
|
|
part = disk_map_sector_rcu(req->rq_disk, req->sector);
|
2008-08-25 12:56:14 +02:00
|
|
|
part_stat_add(cpu, part, sectors[rw], nr_bytes >> 9);
|
|
|
|
part_stat_unlock();
|
2005-11-01 08:35:42 +01:00
|
|
|
}
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
total_bytes = bio_nbytes = 0;
|
|
|
|
while ((bio = req->bio) != NULL) {
|
|
|
|
int nbytes;
|
|
|
|
|
2007-09-27 13:01:25 +02:00
|
|
|
/*
|
|
|
|
* For an empty barrier request, the low level driver must
|
|
|
|
* store a potential error location in ->sector. We pass
|
|
|
|
* that back up in ->bi_sector.
|
|
|
|
*/
|
|
|
|
if (blk_empty_barrier(req))
|
|
|
|
bio->bi_sector = req->sector;
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
if (nr_bytes >= bio->bi_size) {
|
|
|
|
req->bio = bio->bi_next;
|
|
|
|
nbytes = bio->bi_size;
|
2007-09-27 12:46:13 +02:00
|
|
|
req_bio_endio(req, bio, nbytes, error);
|
2005-04-17 00:20:36 +02:00
|
|
|
next_idx = 0;
|
|
|
|
bio_nbytes = 0;
|
|
|
|
} else {
|
|
|
|
int idx = bio->bi_idx + next_idx;
|
|
|
|
|
|
|
|
if (unlikely(bio->bi_idx >= bio->bi_vcnt)) {
|
|
|
|
blk_dump_rq_flags(req, "__end_that");
|
2008-01-31 13:03:55 +01:00
|
|
|
printk(KERN_ERR "%s: bio idx %d >= vcnt %d\n",
|
2008-05-01 13:35:17 +02:00
|
|
|
__func__, bio->bi_idx, bio->bi_vcnt);
|
2005-04-17 00:20:36 +02:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
nbytes = bio_iovec_idx(bio, idx)->bv_len;
|
|
|
|
BIO_BUG_ON(nbytes > bio->bi_size);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* not a complete bvec done
|
|
|
|
*/
|
|
|
|
if (unlikely(nbytes > nr_bytes)) {
|
|
|
|
bio_nbytes += nr_bytes;
|
|
|
|
total_bytes += nr_bytes;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* advance to the next vector
|
|
|
|
*/
|
|
|
|
next_idx++;
|
|
|
|
bio_nbytes += nbytes;
|
|
|
|
}
|
|
|
|
|
|
|
|
total_bytes += nbytes;
|
|
|
|
nr_bytes -= nbytes;
|
|
|
|
|
2008-01-31 13:03:55 +01:00
|
|
|
bio = req->bio;
|
|
|
|
if (bio) {
|
2005-04-17 00:20:36 +02:00
|
|
|
/*
|
|
|
|
* end more in this run, or just return 'not-done'
|
|
|
|
*/
|
|
|
|
if (unlikely(nr_bytes <= 0))
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* completely done
|
|
|
|
*/
|
|
|
|
if (!req->bio)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* if the request wasn't completed, update state
|
|
|
|
*/
|
|
|
|
if (bio_nbytes) {
|
2007-09-27 12:46:13 +02:00
|
|
|
req_bio_endio(req, bio, bio_nbytes, error);
|
2005-04-17 00:20:36 +02:00
|
|
|
bio->bi_idx += next_idx;
|
|
|
|
bio_iovec(bio)->bv_offset += nr_bytes;
|
|
|
|
bio_iovec(bio)->bv_len -= nr_bytes;
|
|
|
|
}
|
|
|
|
|
|
|
|
blk_recalc_rq_sectors(req, total_bytes >> 9);
|
|
|
|
blk_recalc_rq_segments(req);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* queue lock must be held
|
|
|
|
*/
|
2007-12-11 23:53:03 +01:00
|
|
|
static void end_that_request_last(struct request *req, int error)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
struct gendisk *disk = req->rq_disk;
|
2006-01-06 09:49:03 +01:00
|
|
|
|
2007-12-11 23:53:24 +01:00
|
|
|
if (blk_rq_tagged(req))
|
|
|
|
blk_queue_end_tag(req->q, req);
|
|
|
|
|
|
|
|
if (blk_queued_rq(req))
|
|
|
|
blkdev_dequeue_request(req);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
if (unlikely(laptop_mode) && blk_fs_request(req))
|
|
|
|
laptop_io_completion();
|
|
|
|
|
2006-05-23 11:23:49 +02:00
|
|
|
/*
|
|
|
|
* Account IO completion. bar_rq isn't accounted as a normal
|
|
|
|
* IO on queueing nor completion. Accounting the containing
|
|
|
|
* request is enough.
|
|
|
|
*/
|
|
|
|
if (disk && blk_fs_request(req) && req != &req->q->bar_rq) {
|
2005-04-17 00:20:36 +02:00
|
|
|
unsigned long duration = jiffies - req->start_time;
|
2005-11-01 09:26:16 +01:00
|
|
|
const int rw = rq_data_dir(req);
|
2008-09-03 09:03:02 +02:00
|
|
|
struct hd_struct *part;
|
2008-08-25 12:47:21 +02:00
|
|
|
int cpu;
|
2008-09-03 09:03:02 +02:00
|
|
|
|
2008-08-25 12:56:14 +02:00
|
|
|
cpu = part_stat_lock();
|
2008-09-03 09:03:02 +02:00
|
|
|
part = disk_map_sector_rcu(disk, req->sector);
|
2005-11-01 09:26:16 +01:00
|
|
|
|
2008-08-25 12:56:14 +02:00
|
|
|
part_stat_inc(cpu, part, ios[rw]);
|
|
|
|
part_stat_add(cpu, part, ticks[rw], duration);
|
|
|
|
part_round_stats(cpu, part);
|
|
|
|
part_dec_in_flight(part);
|
2008-09-03 09:03:02 +02:00
|
|
|
|
2008-08-25 12:56:14 +02:00
|
|
|
part_stat_unlock();
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
2007-12-11 23:53:24 +01:00
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
if (req->end_io)
|
2006-01-06 09:49:03 +01:00
|
|
|
req->end_io(req, error);
|
2007-12-11 23:53:24 +01:00
|
|
|
else {
|
|
|
|
if (blk_bidi_rq(req))
|
|
|
|
__blk_put_request(req->next_rq->q, req->next_rq);
|
|
|
|
|
2005-04-17 00:20:36 +02:00
|
|
|
__blk_put_request(req->q, req);
|
2007-12-11 23:53:24 +01:00
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2007-09-21 10:41:07 +02:00
|
|
|
static inline void __end_request(struct request *rq, int uptodate,
|
2007-12-11 23:41:54 +01:00
|
|
|
unsigned int nr_bytes)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-12-11 23:41:54 +01:00
|
|
|
int error = 0;
|
|
|
|
|
|
|
|
if (uptodate <= 0)
|
|
|
|
error = uptodate ? uptodate : -EIO;
|
|
|
|
|
|
|
|
__blk_end_request(rq, error, nr_bytes);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|
2007-12-11 23:41:17 +01:00
|
|
|
/**
|
|
|
|
* blk_rq_bytes - Returns bytes left to complete in the entire request
|
2008-02-20 09:01:22 +01:00
|
|
|
* @rq: the request being processed
|
2007-12-11 23:41:17 +01:00
|
|
|
**/
|
|
|
|
unsigned int blk_rq_bytes(struct request *rq)
|
2007-09-21 10:41:07 +02:00
|
|
|
{
|
|
|
|
if (blk_fs_request(rq))
|
|
|
|
return rq->hard_nr_sectors << 9;
|
|
|
|
|
|
|
|
return rq->data_len;
|
|
|
|
}
|
2007-12-11 23:41:17 +01:00
|
|
|
EXPORT_SYMBOL_GPL(blk_rq_bytes);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* blk_rq_cur_bytes - Returns bytes left to complete in the current segment
|
2008-02-20 09:01:22 +01:00
|
|
|
* @rq: the request being processed
|
2007-12-11 23:41:17 +01:00
|
|
|
**/
|
|
|
|
unsigned int blk_rq_cur_bytes(struct request *rq)
|
|
|
|
{
|
|
|
|
if (blk_fs_request(rq))
|
|
|
|
return rq->current_nr_sectors << 9;
|
|
|
|
|
|
|
|
if (rq->bio)
|
|
|
|
return rq->bio->bi_size;
|
|
|
|
|
|
|
|
return rq->data_len;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blk_rq_cur_bytes);
|
2007-09-21 10:41:07 +02:00
|
|
|
|
|
|
|
/**
|
|
|
|
* end_queued_request - end all I/O on a queued request
|
|
|
|
* @rq: the request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @uptodate: error value or %0/%1 uptodate flag
|
2007-09-21 10:41:07 +02:00
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Ends all I/O on a request, and removes it from the block layer queues.
|
2008-08-19 20:13:11 +02:00
|
|
|
* Not suitable for normal I/O completion, unless the driver still has
|
2007-09-21 10:41:07 +02:00
|
|
|
* the request attached to the block layer.
|
|
|
|
*
|
|
|
|
**/
|
|
|
|
void end_queued_request(struct request *rq, int uptodate)
|
|
|
|
{
|
2007-12-11 23:41:54 +01:00
|
|
|
__end_request(rq, uptodate, blk_rq_bytes(rq));
|
2007-09-21 10:41:07 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(end_queued_request);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* end_dequeued_request - end all I/O on a dequeued request
|
|
|
|
* @rq: the request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @uptodate: error value or %0/%1 uptodate flag
|
2007-09-21 10:41:07 +02:00
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Ends all I/O on a request. The request must already have been
|
|
|
|
* dequeued using blkdev_dequeue_request(), as is normally the case
|
|
|
|
* for most drivers.
|
|
|
|
*
|
|
|
|
**/
|
|
|
|
void end_dequeued_request(struct request *rq, int uptodate)
|
|
|
|
{
|
2007-12-11 23:41:54 +01:00
|
|
|
__end_request(rq, uptodate, blk_rq_bytes(rq));
|
2007-09-21 10:41:07 +02:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(end_dequeued_request);
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* end_request - end I/O on the current segment of the request
|
2007-10-19 08:39:28 +02:00
|
|
|
* @req: the request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @uptodate: error value or %0/%1 uptodate flag
|
2007-09-21 10:41:07 +02:00
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Ends I/O on the current segment of a request. If that is the only
|
|
|
|
* remaining segment, the request is also completed and freed.
|
|
|
|
*
|
2008-08-19 20:13:11 +02:00
|
|
|
* This is a remnant of how older block drivers handled I/O completions.
|
|
|
|
* Modern drivers typically end I/O on the full request in one go, unless
|
2007-09-21 10:41:07 +02:00
|
|
|
* they have a residual value to account for. For that case this function
|
|
|
|
* isn't really useful, unless the residual just happens to be the
|
|
|
|
* full current segment. In other words, don't use this function in new
|
|
|
|
* code. Either use end_request_completely(), or the
|
|
|
|
* end_that_request_chunk() (along with end_that_request_last()) for
|
|
|
|
* partial completions.
|
|
|
|
*
|
|
|
|
**/
|
|
|
|
void end_request(struct request *req, int uptodate)
|
|
|
|
{
|
2007-12-11 23:41:54 +01:00
|
|
|
__end_request(req, uptodate, req->hard_cur_sectors << 9);
|
2007-09-21 10:41:07 +02:00
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
EXPORT_SYMBOL(end_request);
|
|
|
|
|
2007-12-11 23:40:30 +01:00
|
|
|
/**
|
2007-12-11 23:51:02 +01:00
|
|
|
* blk_end_io - Generic end_io function to complete a request.
|
|
|
|
* @rq: the request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @error: %0 for success, < %0 for error
|
2007-12-11 23:51:46 +01:00
|
|
|
* @nr_bytes: number of bytes to complete @rq
|
|
|
|
* @bidi_bytes: number of bytes to complete @rq->next_rq
|
2007-12-11 23:51:02 +01:00
|
|
|
* @drv_callback: function called between completion of bios in the request
|
|
|
|
* and completion of the request.
|
2008-08-19 20:13:11 +02:00
|
|
|
* If the callback returns non %0, this helper returns without
|
2007-12-11 23:51:02 +01:00
|
|
|
* completion of the request.
|
2007-12-11 23:40:30 +01:00
|
|
|
*
|
|
|
|
* Description:
|
2007-12-11 23:51:46 +01:00
|
|
|
* Ends I/O on a number of bytes attached to @rq and @rq->next_rq.
|
2007-12-11 23:40:30 +01:00
|
|
|
* If @rq has leftover, sets it up for the next range of segments.
|
|
|
|
*
|
|
|
|
* Return:
|
2008-08-19 20:13:11 +02:00
|
|
|
* %0 - we are done with this request
|
|
|
|
* %1 - this request is not freed yet, it still has pending buffers.
|
2007-12-11 23:40:30 +01:00
|
|
|
**/
|
2008-01-31 12:36:19 +01:00
|
|
|
static int blk_end_io(struct request *rq, int error, unsigned int nr_bytes,
|
|
|
|
unsigned int bidi_bytes,
|
|
|
|
int (drv_callback)(struct request *))
|
2007-12-11 23:40:30 +01:00
|
|
|
{
|
|
|
|
struct request_queue *q = rq->q;
|
|
|
|
unsigned long flags = 0UL;
|
|
|
|
|
2008-08-26 13:34:34 +02:00
|
|
|
if (rq->bio) {
|
2007-12-11 23:53:03 +01:00
|
|
|
if (__end_that_request_first(rq, error, nr_bytes))
|
2007-12-11 23:40:30 +01:00
|
|
|
return 1;
|
2007-12-11 23:51:46 +01:00
|
|
|
|
|
|
|
/* Bidi request must be completed as a whole */
|
|
|
|
if (blk_bidi_rq(rq) &&
|
2007-12-11 23:53:03 +01:00
|
|
|
__end_that_request_first(rq->next_rq, error, bidi_bytes))
|
2007-12-11 23:51:46 +01:00
|
|
|
return 1;
|
2007-12-11 23:40:30 +01:00
|
|
|
}
|
|
|
|
|
2007-12-11 23:51:02 +01:00
|
|
|
/* Special feature for tricky drivers */
|
|
|
|
if (drv_callback && drv_callback(rq))
|
|
|
|
return 1;
|
|
|
|
|
2007-12-11 23:40:30 +01:00
|
|
|
add_disk_randomness(rq->rq_disk);
|
|
|
|
|
|
|
|
spin_lock_irqsave(q->queue_lock, flags);
|
2007-12-11 23:53:24 +01:00
|
|
|
end_that_request_last(rq, error);
|
2007-12-11 23:40:30 +01:00
|
|
|
spin_unlock_irqrestore(q->queue_lock, flags);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2007-12-11 23:51:02 +01:00
|
|
|
|
|
|
|
/**
|
|
|
|
* blk_end_request - Helper function for drivers to complete the request.
|
|
|
|
* @rq: the request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @error: %0 for success, < %0 for error
|
2007-12-11 23:51:02 +01:00
|
|
|
* @nr_bytes: number of bytes to complete
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Ends I/O on a number of bytes attached to @rq.
|
|
|
|
* If @rq has leftover, sets it up for the next range of segments.
|
|
|
|
*
|
|
|
|
* Return:
|
2008-08-19 20:13:11 +02:00
|
|
|
* %0 - we are done with this request
|
|
|
|
* %1 - still buffers pending for this request
|
2007-12-11 23:51:02 +01:00
|
|
|
**/
|
2008-01-31 12:36:19 +01:00
|
|
|
int blk_end_request(struct request *rq, int error, unsigned int nr_bytes)
|
2007-12-11 23:51:02 +01:00
|
|
|
{
|
2007-12-11 23:51:46 +01:00
|
|
|
return blk_end_io(rq, error, nr_bytes, 0, NULL);
|
2007-12-11 23:51:02 +01:00
|
|
|
}
|
2007-12-11 23:40:30 +01:00
|
|
|
EXPORT_SYMBOL_GPL(blk_end_request);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* __blk_end_request - Helper function for drivers to complete the request.
|
|
|
|
* @rq: the request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @error: %0 for success, < %0 for error
|
2007-12-11 23:40:30 +01:00
|
|
|
* @nr_bytes: number of bytes to complete
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Must be called with queue lock held unlike blk_end_request().
|
|
|
|
*
|
|
|
|
* Return:
|
2008-08-19 20:13:11 +02:00
|
|
|
* %0 - we are done with this request
|
|
|
|
* %1 - still buffers pending for this request
|
2007-12-11 23:40:30 +01:00
|
|
|
**/
|
2008-01-31 12:36:19 +01:00
|
|
|
int __blk_end_request(struct request *rq, int error, unsigned int nr_bytes)
|
2007-12-11 23:40:30 +01:00
|
|
|
{
|
2008-08-26 13:34:34 +02:00
|
|
|
if (rq->bio && __end_that_request_first(rq, error, nr_bytes))
|
2008-08-08 11:06:45 +02:00
|
|
|
return 1;
|
2007-12-11 23:40:30 +01:00
|
|
|
|
|
|
|
add_disk_randomness(rq->rq_disk);
|
|
|
|
|
2007-12-11 23:53:24 +01:00
|
|
|
end_that_request_last(rq, error);
|
2007-12-11 23:40:30 +01:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(__blk_end_request);
|
|
|
|
|
2007-12-11 23:51:46 +01:00
|
|
|
/**
|
|
|
|
* blk_end_bidi_request - Helper function for drivers to complete bidi request.
|
|
|
|
* @rq: the bidi request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @error: %0 for success, < %0 for error
|
2007-12-11 23:51:46 +01:00
|
|
|
* @nr_bytes: number of bytes to complete @rq
|
|
|
|
* @bidi_bytes: number of bytes to complete @rq->next_rq
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Ends I/O on a number of bytes attached to @rq and @rq->next_rq.
|
|
|
|
*
|
|
|
|
* Return:
|
2008-08-19 20:13:11 +02:00
|
|
|
* %0 - we are done with this request
|
|
|
|
* %1 - still buffers pending for this request
|
2007-12-11 23:51:46 +01:00
|
|
|
**/
|
2008-01-31 12:36:19 +01:00
|
|
|
int blk_end_bidi_request(struct request *rq, int error, unsigned int nr_bytes,
|
|
|
|
unsigned int bidi_bytes)
|
2007-12-11 23:51:46 +01:00
|
|
|
{
|
|
|
|
return blk_end_io(rq, error, nr_bytes, bidi_bytes, NULL);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blk_end_bidi_request);
|
|
|
|
|
2007-12-11 23:51:02 +01:00
|
|
|
/**
|
|
|
|
* blk_end_request_callback - Special helper function for tricky drivers
|
|
|
|
* @rq: the request being processed
|
2008-08-19 20:13:11 +02:00
|
|
|
* @error: %0 for success, < %0 for error
|
2007-12-11 23:51:02 +01:00
|
|
|
* @nr_bytes: number of bytes to complete
|
|
|
|
* @drv_callback: function called between completion of bios in the request
|
|
|
|
* and completion of the request.
|
2008-08-19 20:13:11 +02:00
|
|
|
* If the callback returns non %0, this helper returns without
|
2007-12-11 23:51:02 +01:00
|
|
|
* completion of the request.
|
|
|
|
*
|
|
|
|
* Description:
|
|
|
|
* Ends I/O on a number of bytes attached to @rq.
|
|
|
|
* If @rq has leftover, sets it up for the next range of segments.
|
|
|
|
*
|
|
|
|
* This special helper function is used only for existing tricky drivers.
|
|
|
|
* (e.g. cdrom_newpc_intr() of ide-cd)
|
|
|
|
* This interface will be removed when such drivers are rewritten.
|
|
|
|
* Don't use this interface in other places anymore.
|
|
|
|
*
|
|
|
|
* Return:
|
2008-08-19 20:13:11 +02:00
|
|
|
* %0 - we are done with this request
|
|
|
|
* %1 - this request is not freed yet.
|
|
|
|
* this request still has pending buffers or
|
|
|
|
* the driver doesn't want to finish this request yet.
|
2007-12-11 23:51:02 +01:00
|
|
|
**/
|
2008-01-31 12:36:19 +01:00
|
|
|
int blk_end_request_callback(struct request *rq, int error,
|
|
|
|
unsigned int nr_bytes,
|
2007-12-11 23:51:02 +01:00
|
|
|
int (drv_callback)(struct request *))
|
|
|
|
{
|
2007-12-11 23:51:46 +01:00
|
|
|
return blk_end_io(rq, error, nr_bytes, 0, drv_callback);
|
2007-12-11 23:51:02 +01:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(blk_end_request_callback);
|
|
|
|
|
2008-01-29 14:53:40 +01:00
|
|
|
void blk_rq_bio_prep(struct request_queue *q, struct request *rq,
|
|
|
|
struct bio *bio)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2008-08-09 17:22:17 +02:00
|
|
|
/* Bit 0 (R/W) is identical in rq->cmd_flags and bio->bi_rw, and
|
|
|
|
we want BIO_RW_AHEAD (bit 1) to imply REQ_FAILFAST (bit 1). */
|
2006-08-10 08:44:47 +02:00
|
|
|
rq->cmd_flags |= (bio->bi_rw & 3);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-08-05 19:01:53 +02:00
|
|
|
if (bio_has_data(bio)) {
|
|
|
|
rq->nr_phys_segments = bio_phys_segments(q, bio);
|
|
|
|
rq->buffer = bio_data(bio);
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
rq->current_nr_sectors = bio_cur_sectors(bio);
|
|
|
|
rq->hard_cur_sectors = rq->current_nr_sectors;
|
|
|
|
rq->hard_nr_sectors = rq->nr_sectors = bio_sectors(bio);
|
2006-12-01 10:40:55 +01:00
|
|
|
rq->data_len = bio->bi_size;
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
rq->bio = rq->biotail = bio;
|
|
|
|
|
2007-08-16 13:31:28 +02:00
|
|
|
if (bio->bi_bdev)
|
|
|
|
rq->rq_disk = bio->bi_bdev->bd_disk;
|
|
|
|
}
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-07-28 13:08:45 +02:00
|
|
|
int kblockd_schedule_work(struct request_queue *q, struct work_struct *work)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
|
|
|
return queue_work(kblockd_workqueue, work);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(kblockd_schedule_work);
|
|
|
|
|
2007-05-09 11:33:56 +02:00
|
|
|
void kblockd_flush_work(struct work_struct *work)
|
2005-04-17 00:20:36 +02:00
|
|
|
{
|
2007-05-09 11:34:22 +02:00
|
|
|
cancel_work_sync(work);
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
2007-05-09 11:33:56 +02:00
|
|
|
EXPORT_SYMBOL(kblockd_flush_work);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
|
|
|
int __init blk_dev_init(void)
|
|
|
|
{
|
|
|
|
kblockd_workqueue = create_workqueue("kblockd");
|
|
|
|
if (!kblockd_workqueue)
|
|
|
|
panic("Failed to create kblockd\n");
|
|
|
|
|
|
|
|
request_cachep = kmem_cache_create("blkdev_requests",
|
2007-07-20 03:11:58 +02:00
|
|
|
sizeof(struct request), 0, SLAB_PANIC, NULL);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-29 14:51:59 +01:00
|
|
|
blk_requestq_cachep = kmem_cache_create("blkdev_queue",
|
2007-07-24 09:28:11 +02:00
|
|
|
sizeof(struct request_queue), 0, SLAB_PANIC, NULL);
|
2005-04-17 00:20:36 +02:00
|
|
|
|
2008-01-24 08:53:35 +01:00
|
|
|
return 0;
|
2005-04-17 00:20:36 +02:00
|
|
|
}
|
|
|
|
|