block: continue ll_rw_blk.c splitup

Adds files for barrier handling, rq execution, io context handling,
mapping data to requests, and queue settings.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
This commit is contained in:
Jens Axboe 2008-01-29 14:53:40 +01:00
parent 8324aa91d1
commit 86db1e2977
8 changed files with 1312 additions and 1249 deletions

View File

@ -2,8 +2,9 @@
# Makefile for the kernel block layer
#
obj-$(CONFIG_BLOCK) := elevator.o blk-core.o blk-tag.o blk-sysfs.o ioctl.o \
genhd.o scsi_ioctl.o
obj-$(CONFIG_BLOCK) := elevator.o blk-core.o blk-tag.o blk-sysfs.o \
blk-barrier.o blk-settings.o blk-ioc.o blk-map.o \
blk-exec.o ioctl.o genhd.o scsi_ioctl.o
obj-$(CONFIG_BLK_DEV_BSG) += bsg.o
obj-$(CONFIG_IOSCHED_NOOP) += noop-iosched.o

319
block/blk-barrier.c Normal file
View File

@ -0,0 +1,319 @@
/*
* Functions related to barrier IO handling
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include "blk.h"
/**
* blk_queue_ordered - does this queue support ordered writes
* @q: the request queue
* @ordered: one of QUEUE_ORDERED_*
* @prepare_flush_fn: rq setup helper for cache flush ordered writes
*
* Description:
* For journalled file systems, doing ordered writes on a commit
* block instead of explicitly doing wait_on_buffer (which is bad
* for performance) can be a big win. Block drivers supporting this
* feature should call this function and indicate so.
*
**/
int blk_queue_ordered(struct request_queue *q, unsigned ordered,
prepare_flush_fn *prepare_flush_fn)
{
if (ordered & (QUEUE_ORDERED_PREFLUSH | QUEUE_ORDERED_POSTFLUSH) &&
prepare_flush_fn == NULL) {
printk(KERN_ERR "blk_queue_ordered: prepare_flush_fn required\n");
return -EINVAL;
}
if (ordered != QUEUE_ORDERED_NONE &&
ordered != QUEUE_ORDERED_DRAIN &&
ordered != QUEUE_ORDERED_DRAIN_FLUSH &&
ordered != QUEUE_ORDERED_DRAIN_FUA &&
ordered != QUEUE_ORDERED_TAG &&
ordered != QUEUE_ORDERED_TAG_FLUSH &&
ordered != QUEUE_ORDERED_TAG_FUA) {
printk(KERN_ERR "blk_queue_ordered: bad value %d\n", ordered);
return -EINVAL;
}
q->ordered = ordered;
q->next_ordered = ordered;
q->prepare_flush_fn = prepare_flush_fn;
return 0;
}
EXPORT_SYMBOL(blk_queue_ordered);
/*
* Cache flushing for ordered writes handling
*/
inline unsigned blk_ordered_cur_seq(struct request_queue *q)
{
if (!q->ordseq)
return 0;
return 1 << ffz(q->ordseq);
}
unsigned blk_ordered_req_seq(struct request *rq)
{
struct request_queue *q = rq->q;
BUG_ON(q->ordseq == 0);
if (rq == &q->pre_flush_rq)
return QUEUE_ORDSEQ_PREFLUSH;
if (rq == &q->bar_rq)
return QUEUE_ORDSEQ_BAR;
if (rq == &q->post_flush_rq)
return QUEUE_ORDSEQ_POSTFLUSH;
/*
* !fs requests don't need to follow barrier ordering. Always
* put them at the front. This fixes the following deadlock.
*
* http://thread.gmane.org/gmane.linux.kernel/537473
*/
if (!blk_fs_request(rq))
return QUEUE_ORDSEQ_DRAIN;
if ((rq->cmd_flags & REQ_ORDERED_COLOR) ==
(q->orig_bar_rq->cmd_flags & REQ_ORDERED_COLOR))
return QUEUE_ORDSEQ_DRAIN;
else
return QUEUE_ORDSEQ_DONE;
}
void blk_ordered_complete_seq(struct request_queue *q, unsigned seq, int error)
{
struct request *rq;
if (error && !q->orderr)
q->orderr = error;
BUG_ON(q->ordseq & seq);
q->ordseq |= seq;
if (blk_ordered_cur_seq(q) != QUEUE_ORDSEQ_DONE)
return;
/*
* Okay, sequence complete.
*/
q->ordseq = 0;
rq = q->orig_bar_rq;
if (__blk_end_request(rq, q->orderr, blk_rq_bytes(rq)))
BUG();
}
static void pre_flush_end_io(struct request *rq, int error)
{
elv_completed_request(rq->q, rq);
blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_PREFLUSH, error);
}
static void bar_end_io(struct request *rq, int error)
{
elv_completed_request(rq->q, rq);
blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_BAR, error);
}
static void post_flush_end_io(struct request *rq, int error)
{
elv_completed_request(rq->q, rq);
blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_POSTFLUSH, error);
}
static void queue_flush(struct request_queue *q, unsigned which)
{
struct request *rq;
rq_end_io_fn *end_io;
if (which == QUEUE_ORDERED_PREFLUSH) {
rq = &q->pre_flush_rq;
end_io = pre_flush_end_io;
} else {
rq = &q->post_flush_rq;
end_io = post_flush_end_io;
}
rq->cmd_flags = REQ_HARDBARRIER;
rq_init(q, rq);
rq->elevator_private = NULL;
rq->elevator_private2 = NULL;
rq->rq_disk = q->bar_rq.rq_disk;
rq->end_io = end_io;
q->prepare_flush_fn(q, rq);
elv_insert(q, rq, ELEVATOR_INSERT_FRONT);
}
static inline struct request *start_ordered(struct request_queue *q,
struct request *rq)
{
q->orderr = 0;
q->ordered = q->next_ordered;
q->ordseq |= QUEUE_ORDSEQ_STARTED;
/*
* Prep proxy barrier request.
*/
blkdev_dequeue_request(rq);
q->orig_bar_rq = rq;
rq = &q->bar_rq;
rq->cmd_flags = 0;
rq_init(q, rq);
if (bio_data_dir(q->orig_bar_rq->bio) == WRITE)
rq->cmd_flags |= REQ_RW;
if (q->ordered & QUEUE_ORDERED_FUA)
rq->cmd_flags |= REQ_FUA;
rq->elevator_private = NULL;
rq->elevator_private2 = NULL;
init_request_from_bio(rq, q->orig_bar_rq->bio);
rq->end_io = bar_end_io;
/*
* Queue ordered sequence. As we stack them at the head, we
* need to queue in reverse order. Note that we rely on that
* no fs request uses ELEVATOR_INSERT_FRONT and thus no fs
* request gets inbetween ordered sequence. If this request is
* an empty barrier, we don't need to do a postflush ever since
* there will be no data written between the pre and post flush.
* Hence a single flush will suffice.
*/
if ((q->ordered & QUEUE_ORDERED_POSTFLUSH) && !blk_empty_barrier(rq))
queue_flush(q, QUEUE_ORDERED_POSTFLUSH);
else
q->ordseq |= QUEUE_ORDSEQ_POSTFLUSH;
elv_insert(q, rq, ELEVATOR_INSERT_FRONT);
if (q->ordered & QUEUE_ORDERED_PREFLUSH) {
queue_flush(q, QUEUE_ORDERED_PREFLUSH);
rq = &q->pre_flush_rq;
} else
q->ordseq |= QUEUE_ORDSEQ_PREFLUSH;
if ((q->ordered & QUEUE_ORDERED_TAG) || q->in_flight == 0)
q->ordseq |= QUEUE_ORDSEQ_DRAIN;
else
rq = NULL;
return rq;
}
int blk_do_ordered(struct request_queue *q, struct request **rqp)
{
struct request *rq = *rqp;
const int is_barrier = blk_fs_request(rq) && blk_barrier_rq(rq);
if (!q->ordseq) {
if (!is_barrier)
return 1;
if (q->next_ordered != QUEUE_ORDERED_NONE) {
*rqp = start_ordered(q, rq);
return 1;
} else {
/*
* This can happen when the queue switches to
* ORDERED_NONE while this request is on it.
*/
blkdev_dequeue_request(rq);
if (__blk_end_request(rq, -EOPNOTSUPP,
blk_rq_bytes(rq)))
BUG();
*rqp = NULL;
return 0;
}
}
/*
* Ordered sequence in progress
*/
/* Special requests are not subject to ordering rules. */
if (!blk_fs_request(rq) &&
rq != &q->pre_flush_rq && rq != &q->post_flush_rq)
return 1;
if (q->ordered & QUEUE_ORDERED_TAG) {
/* Ordered by tag. Blocking the next barrier is enough. */
if (is_barrier && rq != &q->bar_rq)
*rqp = NULL;
} else {
/* Ordered by draining. Wait for turn. */
WARN_ON(blk_ordered_req_seq(rq) < blk_ordered_cur_seq(q));
if (blk_ordered_req_seq(rq) > blk_ordered_cur_seq(q))
*rqp = NULL;
}
return 1;
}
static void bio_end_empty_barrier(struct bio *bio, int err)
{
if (err)
clear_bit(BIO_UPTODATE, &bio->bi_flags);
complete(bio->bi_private);
}
/**
* blkdev_issue_flush - queue a flush
* @bdev: blockdev to issue flush for
* @error_sector: error sector
*
* Description:
* Issue a flush for the block device in question. Caller can supply
* room for storing the error offset in case of a flush error, if they
* wish to. Caller must run wait_for_completion() on its own.
*/
int blkdev_issue_flush(struct block_device *bdev, sector_t *error_sector)
{
DECLARE_COMPLETION_ONSTACK(wait);
struct request_queue *q;
struct bio *bio;
int ret;
if (bdev->bd_disk == NULL)
return -ENXIO;
q = bdev_get_queue(bdev);
if (!q)
return -ENXIO;
bio = bio_alloc(GFP_KERNEL, 0);
if (!bio)
return -ENOMEM;
bio->bi_end_io = bio_end_empty_barrier;
bio->bi_private = &wait;
bio->bi_bdev = bdev;
submit_bio(1 << BIO_RW_BARRIER, bio);
wait_for_completion(&wait);
/*
* The driver must store the error location in ->bi_sector, if
* it supports it. For non-stacked drivers, this should be copied
* from rq->sector.
*/
if (error_sector)
*error_sector = bio->bi_sector;
ret = 0;
if (!bio_flagged(bio, BIO_UPTODATE))
ret = -EIO;
bio_put(bio);
return ret;
}
EXPORT_SYMBOL(blkdev_issue_flush);

File diff suppressed because it is too large Load Diff

105
block/blk-exec.c Normal file
View File

@ -0,0 +1,105 @@
/*
* Functions related to setting various queue properties from drivers
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include "blk.h"
/*
* for max sense size
*/
#include <scsi/scsi_cmnd.h>
/**
* blk_end_sync_rq - executes a completion event on a request
* @rq: request to complete
* @error: end io status of the request
*/
void blk_end_sync_rq(struct request *rq, int error)
{
struct completion *waiting = rq->end_io_data;
rq->end_io_data = NULL;
__blk_put_request(rq->q, rq);
/*
* complete last, if this is a stack request the process (and thus
* the rq pointer) could be invalid right after this complete()
*/
complete(waiting);
}
EXPORT_SYMBOL(blk_end_sync_rq);
/**
* blk_execute_rq_nowait - insert a request into queue for execution
* @q: queue to insert the request in
* @bd_disk: matching gendisk
* @rq: request to insert
* @at_head: insert request at head or tail of queue
* @done: I/O completion handler
*
* Description:
* Insert a fully prepared request at the back of the io scheduler queue
* for execution. Don't wait for completion.
*/
void blk_execute_rq_nowait(struct request_queue *q, struct gendisk *bd_disk,
struct request *rq, int at_head,
rq_end_io_fn *done)
{
int where = at_head ? ELEVATOR_INSERT_FRONT : ELEVATOR_INSERT_BACK;
rq->rq_disk = bd_disk;
rq->cmd_flags |= REQ_NOMERGE;
rq->end_io = done;
WARN_ON(irqs_disabled());
spin_lock_irq(q->queue_lock);
__elv_add_request(q, rq, where, 1);
__generic_unplug_device(q);
spin_unlock_irq(q->queue_lock);
}
EXPORT_SYMBOL_GPL(blk_execute_rq_nowait);
/**
* blk_execute_rq - insert a request into queue for execution
* @q: queue to insert the request in
* @bd_disk: matching gendisk
* @rq: request to insert
* @at_head: insert request at head or tail of queue
*
* Description:
* Insert a fully prepared request at the back of the io scheduler queue
* for execution and wait for completion.
*/
int blk_execute_rq(struct request_queue *q, struct gendisk *bd_disk,
struct request *rq, int at_head)
{
DECLARE_COMPLETION_ONSTACK(wait);
char sense[SCSI_SENSE_BUFFERSIZE];
int err = 0;
/*
* we need an extra reference to the request, so we can look at
* it after io completion
*/
rq->ref_count++;
if (!rq->sense) {
memset(sense, 0, sizeof(sense));
rq->sense = sense;
rq->sense_len = 0;
}
rq->end_io_data = &wait;
blk_execute_rq_nowait(q, bd_disk, rq, at_head, blk_end_sync_rq);
wait_for_completion(&wait);
if (rq->errors)
err = -EIO;
return err;
}
EXPORT_SYMBOL(blk_execute_rq);

194
block/blk-ioc.c Normal file
View File

@ -0,0 +1,194 @@
/*
* Functions related to io context handling
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include <linux/bootmem.h> /* for max_pfn/max_low_pfn */
#include "blk.h"
/*
* For io context allocations
*/
static struct kmem_cache *iocontext_cachep;
static void cfq_dtor(struct io_context *ioc)
{
struct cfq_io_context *cic[1];
int r;
/*
* We don't have a specific key to lookup with, so use the gang
* lookup to just retrieve the first item stored. The cfq exit
* function will iterate the full tree, so any member will do.
*/
r = radix_tree_gang_lookup(&ioc->radix_root, (void **) cic, 0, 1);
if (r > 0)
cic[0]->dtor(ioc);
}
/*
* IO Context helper functions. put_io_context() returns 1 if there are no
* more users of this io context, 0 otherwise.
*/
int put_io_context(struct io_context *ioc)
{
if (ioc == NULL)
return 1;
BUG_ON(atomic_read(&ioc->refcount) == 0);
if (atomic_dec_and_test(&ioc->refcount)) {
rcu_read_lock();
if (ioc->aic && ioc->aic->dtor)
ioc->aic->dtor(ioc->aic);
rcu_read_unlock();
cfq_dtor(ioc);
kmem_cache_free(iocontext_cachep, ioc);
return 1;
}
return 0;
}
EXPORT_SYMBOL(put_io_context);
static void cfq_exit(struct io_context *ioc)
{
struct cfq_io_context *cic[1];
int r;
rcu_read_lock();
/*
* See comment for cfq_dtor()
*/
r = radix_tree_gang_lookup(&ioc->radix_root, (void **) cic, 0, 1);
rcu_read_unlock();
if (r > 0)
cic[0]->exit(ioc);
}
/* Called by the exitting task */
void exit_io_context(void)
{
struct io_context *ioc;
task_lock(current);
ioc = current->io_context;
current->io_context = NULL;
task_unlock(current);
if (atomic_dec_and_test(&ioc->nr_tasks)) {
if (ioc->aic && ioc->aic->exit)
ioc->aic->exit(ioc->aic);
cfq_exit(ioc);
put_io_context(ioc);
}
}
struct io_context *alloc_io_context(gfp_t gfp_flags, int node)
{
struct io_context *ret;
ret = kmem_cache_alloc_node(iocontext_cachep, gfp_flags, node);
if (ret) {
atomic_set(&ret->refcount, 1);
atomic_set(&ret->nr_tasks, 1);
spin_lock_init(&ret->lock);
ret->ioprio_changed = 0;
ret->ioprio = 0;
ret->last_waited = jiffies; /* doesn't matter... */
ret->nr_batch_requests = 0; /* because this is 0 */
ret->aic = NULL;
INIT_RADIX_TREE(&ret->radix_root, GFP_ATOMIC | __GFP_HIGH);
ret->ioc_data = NULL;
}
return ret;
}
/*
* If the current task has no IO context then create one and initialise it.
* Otherwise, return its existing IO context.
*
* This returned IO context doesn't have a specifically elevated refcount,
* but since the current task itself holds a reference, the context can be
* used in general code, so long as it stays within `current` context.
*/
struct io_context *current_io_context(gfp_t gfp_flags, int node)
{
struct task_struct *tsk = current;
struct io_context *ret;
ret = tsk->io_context;
if (likely(ret))
return ret;
ret = alloc_io_context(gfp_flags, node);
if (ret) {
/* make sure set_task_ioprio() sees the settings above */
smp_wmb();
tsk->io_context = ret;
}
return ret;
}
/*
* If the current task has no IO context then create one and initialise it.
* If it does have a context, take a ref on it.
*
* This is always called in the context of the task which submitted the I/O.
*/
struct io_context *get_io_context(gfp_t gfp_flags, int node)
{
struct io_context *ret = NULL;
/*
* Check for unlikely race with exiting task. ioc ref count is
* zero when ioc is being detached.
*/
do {
ret = current_io_context(gfp_flags, node);
if (unlikely(!ret))
break;
} while (!atomic_inc_not_zero(&ret->refcount));
return ret;
}
EXPORT_SYMBOL(get_io_context);
void copy_io_context(struct io_context **pdst, struct io_context **psrc)
{
struct io_context *src = *psrc;
struct io_context *dst = *pdst;
if (src) {
BUG_ON(atomic_read(&src->refcount) == 0);
atomic_inc(&src->refcount);
put_io_context(dst);
*pdst = src;
}
}
EXPORT_SYMBOL(copy_io_context);
void swap_io_context(struct io_context **ioc1, struct io_context **ioc2)
{
struct io_context *temp;
temp = *ioc1;
*ioc1 = *ioc2;
*ioc2 = temp;
}
EXPORT_SYMBOL(swap_io_context);
int __init blk_ioc_init(void)
{
iocontext_cachep = kmem_cache_create("blkdev_ioc",
sizeof(struct io_context), 0, SLAB_PANIC, NULL);
return 0;
}
subsys_initcall(blk_ioc_init);

264
block/blk-map.c Normal file
View File

@ -0,0 +1,264 @@
/*
* Functions related to mapping data to requests
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include "blk.h"
int blk_rq_append_bio(struct request_queue *q, struct request *rq,
struct bio *bio)
{
if (!rq->bio)
blk_rq_bio_prep(q, rq, bio);
else if (!ll_back_merge_fn(q, rq, bio))
return -EINVAL;
else {
rq->biotail->bi_next = bio;
rq->biotail = bio;
rq->data_len += bio->bi_size;
}
return 0;
}
EXPORT_SYMBOL(blk_rq_append_bio);
static int __blk_rq_unmap_user(struct bio *bio)
{
int ret = 0;
if (bio) {
if (bio_flagged(bio, BIO_USER_MAPPED))
bio_unmap_user(bio);
else
ret = bio_uncopy_user(bio);
}
return ret;
}
static int __blk_rq_map_user(struct request_queue *q, struct request *rq,
void __user *ubuf, unsigned int len)
{
unsigned long uaddr;
struct bio *bio, *orig_bio;
int reading, ret;
reading = rq_data_dir(rq) == READ;
/*
* if alignment requirement is satisfied, map in user pages for
* direct dma. else, set up kernel bounce buffers
*/
uaddr = (unsigned long) ubuf;
if (!(uaddr & queue_dma_alignment(q)) && !(len & queue_dma_alignment(q)))
bio = bio_map_user(q, NULL, uaddr, len, reading);
else
bio = bio_copy_user(q, uaddr, len, reading);
if (IS_ERR(bio))
return PTR_ERR(bio);
orig_bio = bio;
blk_queue_bounce(q, &bio);
/*
* We link the bounce buffer in and could have to traverse it
* later so we have to get a ref to prevent it from being freed
*/
bio_get(bio);
ret = blk_rq_append_bio(q, rq, bio);
if (!ret)
return bio->bi_size;
/* if it was boucned we must call the end io function */
bio_endio(bio, 0);
__blk_rq_unmap_user(orig_bio);
bio_put(bio);
return ret;
}
/**
* blk_rq_map_user - map user data to a request, for REQ_BLOCK_PC usage
* @q: request queue where request should be inserted
* @rq: request structure to fill
* @ubuf: the user buffer
* @len: length of user data
*
* Description:
* Data will be mapped directly for zero copy io, if possible. Otherwise
* a kernel bounce buffer is used.
*
* A matching blk_rq_unmap_user() must be issued at the end of io, while
* still in process context.
*
* Note: The mapped bio may need to be bounced through blk_queue_bounce()
* before being submitted to the device, as pages mapped may be out of
* reach. It's the callers responsibility to make sure this happens. The
* original bio must be passed back in to blk_rq_unmap_user() for proper
* unmapping.
*/
int blk_rq_map_user(struct request_queue *q, struct request *rq,
void __user *ubuf, unsigned long len)
{
unsigned long bytes_read = 0;
struct bio *bio = NULL;
int ret;
if (len > (q->max_hw_sectors << 9))
return -EINVAL;
if (!len || !ubuf)
return -EINVAL;
while (bytes_read != len) {
unsigned long map_len, end, start;
map_len = min_t(unsigned long, len - bytes_read, BIO_MAX_SIZE);
end = ((unsigned long)ubuf + map_len + PAGE_SIZE - 1)
>> PAGE_SHIFT;
start = (unsigned long)ubuf >> PAGE_SHIFT;
/*
* A bad offset could cause us to require BIO_MAX_PAGES + 1
* pages. If this happens we just lower the requested
* mapping len by a page so that we can fit
*/
if (end - start > BIO_MAX_PAGES)
map_len -= PAGE_SIZE;
ret = __blk_rq_map_user(q, rq, ubuf, map_len);
if (ret < 0)
goto unmap_rq;
if (!bio)
bio = rq->bio;
bytes_read += ret;
ubuf += ret;
}
rq->buffer = rq->data = NULL;
return 0;
unmap_rq:
blk_rq_unmap_user(bio);
return ret;
}
EXPORT_SYMBOL(blk_rq_map_user);
/**
* blk_rq_map_user_iov - map user data to a request, for REQ_BLOCK_PC usage
* @q: request queue where request should be inserted
* @rq: request to map data to
* @iov: pointer to the iovec
* @iov_count: number of elements in the iovec
* @len: I/O byte count
*
* Description:
* Data will be mapped directly for zero copy io, if possible. Otherwise
* a kernel bounce buffer is used.
*
* A matching blk_rq_unmap_user() must be issued at the end of io, while
* still in process context.
*
* Note: The mapped bio may need to be bounced through blk_queue_bounce()
* before being submitted to the device, as pages mapped may be out of
* reach. It's the callers responsibility to make sure this happens. The
* original bio must be passed back in to blk_rq_unmap_user() for proper
* unmapping.
*/
int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
struct sg_iovec *iov, int iov_count, unsigned int len)
{
struct bio *bio;
if (!iov || iov_count <= 0)
return -EINVAL;
/* we don't allow misaligned data like bio_map_user() does. If the
* user is using sg, they're expected to know the alignment constraints
* and respect them accordingly */
bio = bio_map_user_iov(q, NULL, iov, iov_count, rq_data_dir(rq)== READ);
if (IS_ERR(bio))
return PTR_ERR(bio);
if (bio->bi_size != len) {
bio_endio(bio, 0);
bio_unmap_user(bio);
return -EINVAL;
}
bio_get(bio);
blk_rq_bio_prep(q, rq, bio);
rq->buffer = rq->data = NULL;
return 0;
}
EXPORT_SYMBOL(blk_rq_map_user_iov);
/**
* blk_rq_unmap_user - unmap a request with user data
* @bio: start of bio list
*
* Description:
* Unmap a rq previously mapped by blk_rq_map_user(). The caller must
* supply the original rq->bio from the blk_rq_map_user() return, since
* the io completion may have changed rq->bio.
*/
int blk_rq_unmap_user(struct bio *bio)
{
struct bio *mapped_bio;
int ret = 0, ret2;
while (bio) {
mapped_bio = bio;
if (unlikely(bio_flagged(bio, BIO_BOUNCED)))
mapped_bio = bio->bi_private;
ret2 = __blk_rq_unmap_user(mapped_bio);
if (ret2 && !ret)
ret = ret2;
mapped_bio = bio;
bio = bio->bi_next;
bio_put(mapped_bio);
}
return ret;
}
EXPORT_SYMBOL(blk_rq_unmap_user);
/**
* blk_rq_map_kern - map kernel data to a request, for REQ_BLOCK_PC usage
* @q: request queue where request should be inserted
* @rq: request to fill
* @kbuf: the kernel buffer
* @len: length of user data
* @gfp_mask: memory allocation flags
*/
int blk_rq_map_kern(struct request_queue *q, struct request *rq, void *kbuf,
unsigned int len, gfp_t gfp_mask)
{
struct bio *bio;
if (len > (q->max_hw_sectors << 9))
return -EINVAL;
if (!len || !kbuf)
return -EINVAL;
bio = bio_map_kern(q, kbuf, len, gfp_mask);
if (IS_ERR(bio))
return PTR_ERR(bio);
if (rq_data_dir(rq) == WRITE)
bio->bi_rw |= (1 << BIO_RW);
blk_rq_bio_prep(q, rq, bio);
blk_queue_bounce(q, &rq->bio);
rq->buffer = rq->data = NULL;
return 0;
}
EXPORT_SYMBOL(blk_rq_map_kern);

402
block/blk-settings.c Normal file
View File

@ -0,0 +1,402 @@
/*
* Functions related to setting various queue properties from drivers
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include <linux/bootmem.h> /* for max_pfn/max_low_pfn */
#include "blk.h"
unsigned long blk_max_low_pfn, blk_max_pfn;
EXPORT_SYMBOL(blk_max_low_pfn);
EXPORT_SYMBOL(blk_max_pfn);
/**
* blk_queue_prep_rq - set a prepare_request function for queue
* @q: queue
* @pfn: prepare_request function
*
* It's possible for a queue to register a prepare_request callback which
* is invoked before the request is handed to the request_fn. The goal of
* the function is to prepare a request for I/O, it can be used to build a
* cdb from the request data for instance.
*
*/
void blk_queue_prep_rq(struct request_queue *q, prep_rq_fn *pfn)
{
q->prep_rq_fn = pfn;
}
EXPORT_SYMBOL(blk_queue_prep_rq);
/**
* blk_queue_merge_bvec - set a merge_bvec function for queue
* @q: queue
* @mbfn: merge_bvec_fn
*
* Usually queues have static limitations on the max sectors or segments that
* we can put in a request. Stacking drivers may have some settings that
* are dynamic, and thus we have to query the queue whether it is ok to
* add a new bio_vec to a bio at a given offset or not. If the block device
* has such limitations, it needs to register a merge_bvec_fn to control
* the size of bio's sent to it. Note that a block device *must* allow a
* single page to be added to an empty bio. The block device driver may want
* to use the bio_split() function to deal with these bio's. By default
* no merge_bvec_fn is defined for a queue, and only the fixed limits are
* honored.
*/
void blk_queue_merge_bvec(struct request_queue *q, merge_bvec_fn *mbfn)
{
q->merge_bvec_fn = mbfn;
}
EXPORT_SYMBOL(blk_queue_merge_bvec);
void blk_queue_softirq_done(struct request_queue *q, softirq_done_fn *fn)
{
q->softirq_done_fn = fn;
}
EXPORT_SYMBOL(blk_queue_softirq_done);
/**
* blk_queue_make_request - define an alternate make_request function for a device
* @q: the request queue for the device to be affected
* @mfn: the alternate make_request function
*
* Description:
* The normal way for &struct bios to be passed to a device
* driver is for them to be collected into requests on a request
* queue, and then to allow the device driver to select requests
* off that queue when it is ready. This works well for many block
* devices. However some block devices (typically virtual devices
* such as md or lvm) do not benefit from the processing on the
* request queue, and are served best by having the requests passed
* directly to them. This can be achieved by providing a function
* to blk_queue_make_request().
*
* Caveat:
* The driver that does this *must* be able to deal appropriately
* with buffers in "highmemory". This can be accomplished by either calling
* __bio_kmap_atomic() to get a temporary kernel mapping, or by calling
* blk_queue_bounce() to create a buffer in normal memory.
**/
void blk_queue_make_request(struct request_queue * q, make_request_fn * mfn)
{
/*
* set defaults
*/
q->nr_requests = BLKDEV_MAX_RQ;
blk_queue_max_phys_segments(q, MAX_PHYS_SEGMENTS);
blk_queue_max_hw_segments(q, MAX_HW_SEGMENTS);
q->make_request_fn = mfn;
q->backing_dev_info.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
q->backing_dev_info.state = 0;
q->backing_dev_info.capabilities = BDI_CAP_MAP_COPY;
blk_queue_max_sectors(q, SAFE_MAX_SECTORS);
blk_queue_hardsect_size(q, 512);
blk_queue_dma_alignment(q, 511);
blk_queue_congestion_threshold(q);
q->nr_batching = BLK_BATCH_REQ;
q->unplug_thresh = 4; /* hmm */
q->unplug_delay = (3 * HZ) / 1000; /* 3 milliseconds */
if (q->unplug_delay == 0)
q->unplug_delay = 1;
INIT_WORK(&q->unplug_work, blk_unplug_work);
q->unplug_timer.function = blk_unplug_timeout;
q->unplug_timer.data = (unsigned long)q;
/*
* by default assume old behaviour and bounce for any highmem page
*/
blk_queue_bounce_limit(q, BLK_BOUNCE_HIGH);
}
EXPORT_SYMBOL(blk_queue_make_request);
/**
* blk_queue_bounce_limit - set bounce buffer limit for queue
* @q: the request queue for the device
* @dma_addr: bus address limit
*
* Description:
* Different hardware can have different requirements as to what pages
* it can do I/O directly to. A low level driver can call
* blk_queue_bounce_limit to have lower memory pages allocated as bounce
* buffers for doing I/O to pages residing above @page.
**/
void blk_queue_bounce_limit(struct request_queue *q, u64 dma_addr)
{
unsigned long bounce_pfn = dma_addr >> PAGE_SHIFT;
int dma = 0;
q->bounce_gfp = GFP_NOIO;
#if BITS_PER_LONG == 64
/* Assume anything <= 4GB can be handled by IOMMU.
Actually some IOMMUs can handle everything, but I don't
know of a way to test this here. */
if (bounce_pfn < (min_t(u64,0xffffffff,BLK_BOUNCE_HIGH) >> PAGE_SHIFT))
dma = 1;
q->bounce_pfn = max_low_pfn;
#else
if (bounce_pfn < blk_max_low_pfn)
dma = 1;
q->bounce_pfn = bounce_pfn;
#endif
if (dma) {
init_emergency_isa_pool();
q->bounce_gfp = GFP_NOIO | GFP_DMA;
q->bounce_pfn = bounce_pfn;
}
}
EXPORT_SYMBOL(blk_queue_bounce_limit);
/**
* blk_queue_max_sectors - set max sectors for a request for this queue
* @q: the request queue for the device
* @max_sectors: max sectors in the usual 512b unit
*
* Description:
* Enables a low level driver to set an upper limit on the size of
* received requests.
**/
void blk_queue_max_sectors(struct request_queue *q, unsigned int max_sectors)
{
if ((max_sectors << 9) < PAGE_CACHE_SIZE) {
max_sectors = 1 << (PAGE_CACHE_SHIFT - 9);
printk("%s: set to minimum %d\n", __FUNCTION__, max_sectors);
}
if (BLK_DEF_MAX_SECTORS > max_sectors)
q->max_hw_sectors = q->max_sectors = max_sectors;
else {
q->max_sectors = BLK_DEF_MAX_SECTORS;
q->max_hw_sectors = max_sectors;
}
}
EXPORT_SYMBOL(blk_queue_max_sectors);
/**
* blk_queue_max_phys_segments - set max phys segments for a request for this queue
* @q: the request queue for the device
* @max_segments: max number of segments
*
* Description:
* Enables a low level driver to set an upper limit on the number of
* physical data segments in a request. This would be the largest sized
* scatter list the driver could handle.
**/
void blk_queue_max_phys_segments(struct request_queue *q,
unsigned short max_segments)
{
if (!max_segments) {
max_segments = 1;
printk("%s: set to minimum %d\n", __FUNCTION__, max_segments);
}
q->max_phys_segments = max_segments;
}
EXPORT_SYMBOL(blk_queue_max_phys_segments);
/**
* blk_queue_max_hw_segments - set max hw segments for a request for this queue
* @q: the request queue for the device
* @max_segments: max number of segments
*
* Description:
* Enables a low level driver to set an upper limit on the number of
* hw data segments in a request. This would be the largest number of
* address/length pairs the host adapter can actually give as once
* to the device.
**/
void blk_queue_max_hw_segments(struct request_queue *q,
unsigned short max_segments)
{
if (!max_segments) {
max_segments = 1;
printk("%s: set to minimum %d\n", __FUNCTION__, max_segments);
}
q->max_hw_segments = max_segments;
}
EXPORT_SYMBOL(blk_queue_max_hw_segments);
/**
* blk_queue_max_segment_size - set max segment size for blk_rq_map_sg
* @q: the request queue for the device
* @max_size: max size of segment in bytes
*
* Description:
* Enables a low level driver to set an upper limit on the size of a
* coalesced segment
**/
void blk_queue_max_segment_size(struct request_queue *q, unsigned int max_size)
{
if (max_size < PAGE_CACHE_SIZE) {
max_size = PAGE_CACHE_SIZE;
printk("%s: set to minimum %d\n", __FUNCTION__, max_size);
}
q->max_segment_size = max_size;
}
EXPORT_SYMBOL(blk_queue_max_segment_size);
/**
* blk_queue_hardsect_size - set hardware sector size for the queue
* @q: the request queue for the device
* @size: the hardware sector size, in bytes
*
* Description:
* This should typically be set to the lowest possible sector size
* that the hardware can operate on (possible without reverting to
* even internal read-modify-write operations). Usually the default
* of 512 covers most hardware.
**/
void blk_queue_hardsect_size(struct request_queue *q, unsigned short size)
{
q->hardsect_size = size;
}
EXPORT_SYMBOL(blk_queue_hardsect_size);
/*
* Returns the minimum that is _not_ zero, unless both are zero.
*/
#define min_not_zero(l, r) (l == 0) ? r : ((r == 0) ? l : min(l, r))
/**
* blk_queue_stack_limits - inherit underlying queue limits for stacked drivers
* @t: the stacking driver (top)
* @b: the underlying device (bottom)
**/
void blk_queue_stack_limits(struct request_queue *t, struct request_queue *b)
{
/* zero is "infinity" */
t->max_sectors = min_not_zero(t->max_sectors,b->max_sectors);
t->max_hw_sectors = min_not_zero(t->max_hw_sectors,b->max_hw_sectors);
t->max_phys_segments = min(t->max_phys_segments,b->max_phys_segments);
t->max_hw_segments = min(t->max_hw_segments,b->max_hw_segments);
t->max_segment_size = min(t->max_segment_size,b->max_segment_size);
t->hardsect_size = max(t->hardsect_size,b->hardsect_size);
if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags))
clear_bit(QUEUE_FLAG_CLUSTER, &t->queue_flags);
}
EXPORT_SYMBOL(blk_queue_stack_limits);
/**
* blk_queue_dma_drain - Set up a drain buffer for excess dma.
*
* @q: the request queue for the device
* @buf: physically contiguous buffer
* @size: size of the buffer in bytes
*
* Some devices have excess DMA problems and can't simply discard (or
* zero fill) the unwanted piece of the transfer. They have to have a
* real area of memory to transfer it into. The use case for this is
* ATAPI devices in DMA mode. If the packet command causes a transfer
* bigger than the transfer size some HBAs will lock up if there
* aren't DMA elements to contain the excess transfer. What this API
* does is adjust the queue so that the buf is always appended
* silently to the scatterlist.
*
* Note: This routine adjusts max_hw_segments to make room for
* appending the drain buffer. If you call
* blk_queue_max_hw_segments() or blk_queue_max_phys_segments() after
* calling this routine, you must set the limit to one fewer than your
* device can support otherwise there won't be room for the drain
* buffer.
*/
int blk_queue_dma_drain(struct request_queue *q, void *buf,
unsigned int size)
{
if (q->max_hw_segments < 2 || q->max_phys_segments < 2)
return -EINVAL;
/* make room for appending the drain */
--q->max_hw_segments;
--q->max_phys_segments;
q->dma_drain_buffer = buf;
q->dma_drain_size = size;
return 0;
}
EXPORT_SYMBOL_GPL(blk_queue_dma_drain);
/**
* blk_queue_segment_boundary - set boundary rules for segment merging
* @q: the request queue for the device
* @mask: the memory boundary mask
**/
void blk_queue_segment_boundary(struct request_queue *q, unsigned long mask)
{
if (mask < PAGE_CACHE_SIZE - 1) {
mask = PAGE_CACHE_SIZE - 1;
printk("%s: set to minimum %lx\n", __FUNCTION__, mask);
}
q->seg_boundary_mask = mask;
}
EXPORT_SYMBOL(blk_queue_segment_boundary);
/**
* blk_queue_dma_alignment - set dma length and memory alignment
* @q: the request queue for the device
* @mask: alignment mask
*
* description:
* set required memory and length aligment for direct dma transactions.
* this is used when buiding direct io requests for the queue.
*
**/
void blk_queue_dma_alignment(struct request_queue *q, int mask)
{
q->dma_alignment = mask;
}
EXPORT_SYMBOL(blk_queue_dma_alignment);
/**
* blk_queue_update_dma_alignment - update dma length and memory alignment
* @q: the request queue for the device
* @mask: alignment mask
*
* description:
* update required memory and length aligment for direct dma transactions.
* If the requested alignment is larger than the current alignment, then
* the current queue alignment is updated to the new value, otherwise it
* is left alone. The design of this is to allow multiple objects
* (driver, device, transport etc) to set their respective
* alignments without having them interfere.
*
**/
void blk_queue_update_dma_alignment(struct request_queue *q, int mask)
{
BUG_ON(mask > PAGE_SIZE);
if (mask > q->dma_alignment)
q->dma_alignment = mask;
}
EXPORT_SYMBOL(blk_queue_update_dma_alignment);
int __init blk_settings_init(void)
{
blk_max_low_pfn = max_low_pfn - 1;
blk_max_pfn = max_pfn - 1;
return 0;
}
subsys_initcall(blk_settings_init);

View File

@ -1,11 +1,28 @@
#ifndef BLK_INTERNAL_H
#define BLK_INTERNAL_H
/* Amount of time in which a process may batch requests */
#define BLK_BATCH_TIME (HZ/50UL)
/* Number of requests a "batching" process may submit */
#define BLK_BATCH_REQ 32
extern struct kmem_cache *blk_requestq_cachep;
extern struct kobj_type blk_queue_ktype;
void rq_init(struct request_queue *q, struct request *rq);
void init_request_from_bio(struct request *req, struct bio *bio);
void blk_rq_bio_prep(struct request_queue *q, struct request *rq,
struct bio *bio);
int ll_back_merge_fn(struct request_queue *q, struct request *req,
struct bio *bio);
void __blk_queue_free_tags(struct request_queue *q);
void blk_unplug_work(struct work_struct *work);
void blk_unplug_timeout(unsigned long data);
struct io_context *current_io_context(gfp_t gfp_flags, int node);
void blk_queue_congestion_threshold(struct request_queue *q);
/*