/*
* Anticipatory & deadline i/o scheduler.
*
- * Copyright (C) 2002 Jens Axboe <axboe@suse.de>
+ * Copyright (C) 2002 Jens Axboe <axboe@kernel.dk>
* Nick Piggin <nickpiggin@yahoo.com.au>
*
*/
#include <linux/rbtree.h>
#include <linux/interrupt.h>
-#define REQ_SYNC 1
-#define REQ_ASYNC 0
-
/*
* See Documentation/block/as-iosched.txt
*/
struct rb_root sort_list[2];
struct list_head fifo_list[2];
- struct as_rq *next_arq[2]; /* next in sort order */
- sector_t last_sector[2]; /* last REQ_SYNC & REQ_ASYNC sectors */
+ struct request *next_rq[2]; /* next in sort order */
+ sector_t last_sector[2]; /* last SYNC & ASYNC sectors */
unsigned long exit_prob; /* probability a task will exit while
being waited on */
unsigned long last_check_fifo[2];
int changed_batch; /* 1: waiting for old batch to end */
int new_batch; /* 1: waiting on first read complete */
- int batch_data_dir; /* current batch REQ_SYNC / REQ_ASYNC */
+ int batch_data_dir; /* current batch SYNC / ASYNC */
int write_batch_count; /* max # of reqs in a write batch */
int current_write_count; /* how many requests left this batch */
int write_batch_idled; /* has the write batch gone idle? */
- mempool_t *arq_pool;
enum anticipation_status antic_status;
unsigned long antic_start; /* jiffies: when it started */
AS_RQ_POSTSCHED, /* when they shouldn't be */
};
-struct as_rq {
- struct request *request;
-
- struct io_context *io_context; /* The submitting task */
-
- enum arq_state state;
-};
-
-#define RQ_DATA(rq) ((struct as_rq *) (rq)->elevator_private)
+#define RQ_IOC(rq) ((struct io_context *) (rq)->elevator_private)
+#define RQ_STATE(rq) ((enum arq_state)(rq)->elevator_private2)
+#define RQ_SET_STATE(rq, state) ((rq)->elevator_private2 = (void *) state)
-static kmem_cache_t *arq_pool;
-
-static atomic_t ioc_count = ATOMIC_INIT(0);
+static DEFINE_PER_CPU(unsigned long, as_ioc_count);
static struct completion *ioc_gone;
+static DEFINE_SPINLOCK(ioc_gone_lock);
-static void as_move_to_dispatch(struct as_data *ad, struct as_rq *arq);
+static void as_move_to_dispatch(struct as_data *ad, struct request *rq);
static void as_antic_stop(struct as_data *ad);
/*
static void free_as_io_context(struct as_io_context *aic)
{
kfree(aic);
- if (atomic_dec_and_test(&ioc_count) && ioc_gone)
- complete(ioc_gone);
+ elv_ioc_count_dec(as_ioc_count);
+ if (ioc_gone) {
+ /*
+ * AS scheduler is exiting, grab exit lock and check
+ * the pending io context count. If it hits zero,
+ * complete ioc_gone and set it back to NULL.
+ */
+ spin_lock(&ioc_gone_lock);
+ if (ioc_gone && !elv_ioc_count_read(as_ioc_count)) {
+ complete(ioc_gone);
+ ioc_gone = NULL;
+ }
+ spin_unlock(&ioc_gone_lock);
+ }
}
static void as_trim(struct io_context *ioc)
{
+ spin_lock_irq(&ioc->lock);
if (ioc->aic)
free_as_io_context(ioc->aic);
ioc->aic = NULL;
+ spin_unlock_irq(&ioc->lock);
}
/* Called when the task exits */
ret->seek_total = 0;
ret->seek_samples = 0;
ret->seek_mean = 0;
- atomic_inc(&ioc_count);
+ elv_ioc_count_inc(as_ioc_count);
}
return ret;
* If the current task has no AS IO context then create one and initialise it.
* Then take a ref on the task's io context and return it.
*/
-static struct io_context *as_get_io_context(void)
+static struct io_context *as_get_io_context(int node)
{
- struct io_context *ioc = get_io_context(GFP_ATOMIC);
+ struct io_context *ioc = get_io_context(GFP_ATOMIC, node);
if (ioc && !ioc->aic) {
ioc->aic = alloc_as_io_context();
if (!ioc->aic) {
return ioc;
}
-static void as_put_io_context(struct as_rq *arq)
+static void as_put_io_context(struct request *rq)
{
struct as_io_context *aic;
- if (unlikely(!arq->io_context))
+ if (unlikely(!RQ_IOC(rq)))
return;
- aic = arq->io_context->aic;
+ aic = RQ_IOC(rq)->aic;
- if (rq_is_sync(arq->request) && aic) {
- spin_lock(&aic->lock);
+ if (rq_is_sync(rq) && aic) {
+ unsigned long flags;
+
+ spin_lock_irqsave(&aic->lock, flags);
set_bit(AS_TASK_IORUNNING, &aic->state);
aic->last_end_request = jiffies;
- spin_unlock(&aic->lock);
+ spin_unlock_irqrestore(&aic->lock, flags);
}
- put_io_context(arq->io_context);
+ put_io_context(RQ_IOC(rq));
}
/*
*/
#define RQ_RB_ROOT(ad, rq) (&(ad)->sort_list[rq_is_sync((rq))])
-static void as_add_arq_rb(struct as_data *ad, struct request *rq)
+static void as_add_rq_rb(struct as_data *ad, struct request *rq)
{
struct request *alias;
while ((unlikely(alias = elv_rb_add(RQ_RB_ROOT(ad, rq), rq)))) {
- as_move_to_dispatch(ad, RQ_DATA(alias));
+ as_move_to_dispatch(ad, alias);
as_antic_stop(ad);
}
}
-static inline void as_del_arq_rb(struct as_data *ad, struct request *rq)
+static inline void as_del_rq_rb(struct as_data *ad, struct request *rq)
{
elv_rb_del(RQ_RB_ROOT(ad, rq), rq);
}
* as_choose_req selects the preferred one of two requests of the same data_dir
* ignoring time - eg. timeouts, which is the job of as_dispatch_request
*/
-static struct as_rq *
-as_choose_req(struct as_data *ad, struct as_rq *arq1, struct as_rq *arq2)
+static struct request *
+as_choose_req(struct as_data *ad, struct request *rq1, struct request *rq2)
{
int data_dir;
sector_t last, s1, s2, d1, d2;
int r1_wrap=0, r2_wrap=0; /* requests are behind the disk head */
const sector_t maxback = MAXBACK;
- if (arq1 == NULL || arq1 == arq2)
- return arq2;
- if (arq2 == NULL)
- return arq1;
+ if (rq1 == NULL || rq1 == rq2)
+ return rq2;
+ if (rq2 == NULL)
+ return rq1;
- data_dir = rq_is_sync(arq1->request);
+ data_dir = rq_is_sync(rq1);
last = ad->last_sector[data_dir];
- s1 = arq1->request->sector;
- s2 = arq2->request->sector;
+ s1 = blk_rq_pos(rq1);
+ s2 = blk_rq_pos(rq2);
- BUG_ON(data_dir != rq_is_sync(arq2->request));
+ BUG_ON(data_dir != rq_is_sync(rq2));
/*
* Strict one way elevator _except_ in the case where we allow
/* Found required data */
if (!r1_wrap && r2_wrap)
- return arq1;
+ return rq1;
else if (!r2_wrap && r1_wrap)
- return arq2;
+ return rq2;
else if (r1_wrap && r2_wrap) {
/* both behind the head */
if (s1 <= s2)
- return arq1;
+ return rq1;
else
- return arq2;
+ return rq2;
}
/* Both requests in front of the head */
if (d1 < d2)
- return arq1;
+ return rq1;
else if (d2 < d1)
- return arq2;
+ return rq2;
else {
if (s1 >= s2)
- return arq1;
+ return rq1;
else
- return arq2;
+ return rq2;
}
}
/*
- * as_find_next_arq finds the next request after @prev in elevator order.
+ * as_find_next_rq finds the next request after @prev in elevator order.
* this with as_choose_req form the basis for how the scheduler chooses
* what request to process next. Anticipation works on top of this.
*/
-static struct as_rq *as_find_next_arq(struct as_data *ad, struct as_rq *arq)
+static struct request *
+as_find_next_rq(struct as_data *ad, struct request *last)
{
- struct request *last = arq->request;
struct rb_node *rbnext = rb_next(&last->rb_node);
struct rb_node *rbprev = rb_prev(&last->rb_node);
- struct as_rq *next = NULL, *prev = NULL;
+ struct request *next = NULL, *prev = NULL;
BUG_ON(RB_EMPTY_NODE(&last->rb_node));
if (rbprev)
- prev = RQ_DATA(rb_entry_rq(rbprev));
+ prev = rb_entry_rq(rbprev);
if (rbnext)
- next = RQ_DATA(rb_entry_rq(rbnext));
+ next = rb_entry_rq(rbnext);
else {
const int data_dir = rq_is_sync(last);
rbnext = rb_first(&ad->sort_list[data_dir]);
if (rbnext && rbnext != &last->rb_node)
- next = RQ_DATA(rb_entry_rq(rbnext));
+ next = rb_entry_rq(rbnext);
}
return as_choose_req(ad, next, prev);
del_timer(&ad->antic_timer);
ad->antic_status = ANTIC_FINISHED;
/* see as_work_handler */
- kblockd_schedule_work(&ad->antic_work);
+ kblockd_schedule_work(ad->q, &ad->antic_work);
}
}
spin_lock_irqsave(q->queue_lock, flags);
if (ad->antic_status == ANTIC_WAIT_REQ
|| ad->antic_status == ANTIC_WAIT_NEXT) {
- struct as_io_context *aic = ad->io_context->aic;
+ struct as_io_context *aic;
+ spin_lock(&ad->io_context->lock);
+ aic = ad->io_context->aic;
ad->antic_status = ANTIC_FINISHED;
- kblockd_schedule_work(&ad->antic_work);
+ kblockd_schedule_work(q, &ad->antic_work);
if (aic->ttime_samples == 0) {
/* process anticipated on has exited or timed out*/
/* process not "saved" by a cooperating request */
ad->exit_no_coop = (7*ad->exit_no_coop + 256)/8;
}
+ spin_unlock(&ad->io_context->lock);
}
spin_unlock_irqrestore(q->queue_lock, flags);
}
if (aic == NULL)
return;
- if (data_dir == REQ_SYNC) {
+ if (data_dir == BLK_RW_SYNC) {
unsigned long in_flight = atomic_read(&aic->nr_queued)
+ atomic_read(&aic->nr_dispatched);
spin_lock(&aic->lock);
as_update_thinktime(ad, aic, thinktime);
/* Calculate read -> read seek distance */
- if (aic->last_request_pos < rq->sector)
- seek_dist = rq->sector - aic->last_request_pos;
+ if (aic->last_request_pos < blk_rq_pos(rq))
+ seek_dist = blk_rq_pos(rq) -
+ aic->last_request_pos;
else
- seek_dist = aic->last_request_pos - rq->sector;
+ seek_dist = aic->last_request_pos -
+ blk_rq_pos(rq);
as_update_seekdist(ad, aic, seek_dist);
}
- aic->last_request_pos = rq->sector + rq->nr_sectors;
+ aic->last_request_pos = blk_rq_pos(rq) + blk_rq_sectors(rq);
set_bit(AS_TASK_IOSTARTED, &aic->state);
spin_unlock(&aic->lock);
}
* previous one issued.
*/
static int as_close_req(struct as_data *ad, struct as_io_context *aic,
- struct as_rq *arq)
+ struct request *rq)
{
- unsigned long delay; /* milliseconds */
+ unsigned long delay; /* jiffies */
sector_t last = ad->last_sector[ad->batch_data_dir];
- sector_t next = arq->request->sector;
+ sector_t next = blk_rq_pos(rq);
sector_t delta; /* acceptable close offset (in sectors) */
sector_t s;
if (ad->antic_status == ANTIC_OFF || !ad->ioc_finished)
delay = 0;
else
- delay = ((jiffies - ad->antic_start) * 1000) / HZ;
+ delay = jiffies - ad->antic_start;
if (delay == 0)
delta = 8192;
- else if (delay <= 20 && delay <= ad->antic_expire)
+ else if (delay <= (20 * HZ / 1000) && delay <= ad->antic_expire)
delta = 8192 << delay;
else
return 1;
*
* If this task has queued some other IO, do not enter enticipation.
*/
-static int as_can_break_anticipation(struct as_data *ad, struct as_rq *arq)
+static int as_can_break_anticipation(struct as_data *ad, struct request *rq)
{
struct io_context *ioc;
struct as_io_context *aic;
ioc = ad->io_context;
BUG_ON(!ioc);
+ spin_lock(&ioc->lock);
- if (arq && ioc == arq->io_context) {
+ if (rq && ioc == RQ_IOC(rq)) {
/* request from same process */
+ spin_unlock(&ioc->lock);
return 1;
}
* In this situation status should really be FINISHED,
* however the timer hasn't had the chance to run yet.
*/
+ spin_unlock(&ioc->lock);
return 1;
}
aic = ioc->aic;
- if (!aic)
+ if (!aic) {
+ spin_unlock(&ioc->lock);
return 0;
+ }
if (atomic_read(&aic->nr_queued) > 0) {
/* process has more requests queued */
+ spin_unlock(&ioc->lock);
return 1;
}
if (atomic_read(&aic->nr_dispatched) > 0) {
/* process has more requests dispatched */
+ spin_unlock(&ioc->lock);
return 1;
}
- if (arq && rq_is_sync(arq->request) && as_close_req(ad, aic, arq)) {
+ if (rq && rq_is_sync(rq) && as_close_req(ad, aic, rq)) {
/*
* Found a close request that is not one of ours.
*
ad->exit_no_coop = (7*ad->exit_no_coop)/8;
}
- as_update_iohist(ad, aic, arq->request);
+ as_update_iohist(ad, aic, rq);
+ spin_unlock(&ioc->lock);
return 1;
}
if (aic->ttime_samples == 0)
ad->exit_prob = (7*ad->exit_prob + 256)/8;
- if (ad->exit_no_coop > 128)
+ if (ad->exit_no_coop > 128) {
+ spin_unlock(&ioc->lock);
return 1;
+ }
}
if (aic->ttime_samples == 0) {
- if (ad->new_ttime_mean > ad->antic_expire)
+ if (ad->new_ttime_mean > ad->antic_expire) {
+ spin_unlock(&ioc->lock);
return 1;
- if (ad->exit_prob * ad->exit_no_coop > 128*256)
+ }
+ if (ad->exit_prob * ad->exit_no_coop > 128*256) {
+ spin_unlock(&ioc->lock);
return 1;
+ }
} else if (aic->ttime_mean > ad->antic_expire) {
/* the process thinks too much between requests */
+ spin_unlock(&ioc->lock);
return 1;
}
-
+ spin_unlock(&ioc->lock);
return 0;
}
/*
- * as_can_anticipate indicates whether we should either run arq
+ * as_can_anticipate indicates whether we should either run rq
* or keep anticipating a better request.
*/
-static int as_can_anticipate(struct as_data *ad, struct as_rq *arq)
+static int as_can_anticipate(struct as_data *ad, struct request *rq)
{
+#if 0 /* disable for now, we need to check tag level as well */
+ /*
+ * SSD device without seek penalty, disable idling
+ */
+ if (blk_queue_nonrot(ad->q)) axman
+ return 0;
+#endif
+
if (!ad->io_context)
/*
* Last request submitted was a write
*/
return 0;
- if (as_can_break_anticipation(ad, arq))
+ if (as_can_break_anticipation(ad, rq))
/*
* This request is a good candidate. Don't keep anticipating,
* run it.
}
/*
- * as_update_arq must be called whenever a request (arq) is added to
+ * as_update_rq must be called whenever a request (rq) is added to
* the sort_list. This function keeps caches up to date, and checks if the
* request might be one we are "anticipating"
*/
-static void as_update_arq(struct as_data *ad, struct as_rq *arq)
+static void as_update_rq(struct as_data *ad, struct request *rq)
{
- const int data_dir = rq_is_sync(arq->request);
+ const int data_dir = rq_is_sync(rq);
- /* keep the next_arq cache up to date */
- ad->next_arq[data_dir] = as_choose_req(ad, arq, ad->next_arq[data_dir]);
+ /* keep the next_rq cache up to date */
+ ad->next_rq[data_dir] = as_choose_req(ad, rq, ad->next_rq[data_dir]);
/*
* have we been anticipating this request?
*/
if (ad->antic_status == ANTIC_WAIT_REQ
|| ad->antic_status == ANTIC_WAIT_NEXT) {
- if (as_can_break_anticipation(ad, arq))
+ if (as_can_break_anticipation(ad, rq))
as_antic_stop(ad);
}
}
*/
static void update_write_batch(struct as_data *ad)
{
- unsigned long batch = ad->batch_expire[REQ_ASYNC];
+ unsigned long batch = ad->batch_expire[BLK_RW_ASYNC];
long write_time;
write_time = (jiffies - ad->current_batch_expires) + batch;
* as_completed_request is to be called when a request has completed and
* returned something to the requesting process, be it an error or data.
*/
-static void as_completed_request(request_queue_t *q, struct request *rq)
+static void as_completed_request(struct request_queue *q, struct request *rq)
{
struct as_data *ad = q->elevator->elevator_data;
- struct as_rq *arq = RQ_DATA(rq);
WARN_ON(!list_empty(&rq->queuelist));
- if (arq->state != AS_RQ_REMOVED) {
- printk("arq->state %d\n", arq->state);
- WARN_ON(1);
+ if (RQ_STATE(rq) != AS_RQ_REMOVED) {
+ WARN(1, "rq->state %d\n", RQ_STATE(rq));
goto out;
}
if (ad->changed_batch && ad->nr_dispatched == 1) {
- kblockd_schedule_work(&ad->antic_work);
+ ad->current_batch_expires = jiffies +
+ ad->batch_expire[ad->batch_data_dir];
+ kblockd_schedule_work(q, &ad->antic_work);
ad->changed_batch = 0;
- if (ad->batch_data_dir == REQ_SYNC)
+ if (ad->batch_data_dir == BLK_RW_SYNC)
ad->new_batch = 1;
}
WARN_ON(ad->nr_dispatched == 0);
if (ad->new_batch && ad->batch_data_dir == rq_is_sync(rq)) {
update_write_batch(ad);
ad->current_batch_expires = jiffies +
- ad->batch_expire[REQ_SYNC];
+ ad->batch_expire[BLK_RW_SYNC];
ad->new_batch = 0;
}
- if (ad->io_context == arq->io_context && ad->io_context) {
+ if (ad->io_context == RQ_IOC(rq) && ad->io_context) {
ad->antic_start = jiffies;
ad->ioc_finished = 1;
if (ad->antic_status == ANTIC_WAIT_REQ) {
}
}
- as_put_io_context(arq);
+ as_put_io_context(rq);
out:
- arq->state = AS_RQ_POSTSCHED;
+ RQ_SET_STATE(rq, AS_RQ_POSTSCHED);
}
/*
* reference unless it replaces the request at somepart of the elevator
* (ie. the dispatch queue)
*/
-static void as_remove_queued_request(request_queue_t *q, struct request *rq)
+static void as_remove_queued_request(struct request_queue *q,
+ struct request *rq)
{
- struct as_rq *arq = RQ_DATA(rq);
const int data_dir = rq_is_sync(rq);
struct as_data *ad = q->elevator->elevator_data;
+ struct io_context *ioc;
- WARN_ON(arq->state != AS_RQ_QUEUED);
+ WARN_ON(RQ_STATE(rq) != AS_RQ_QUEUED);
- if (arq->io_context && arq->io_context->aic) {
- BUG_ON(!atomic_read(&arq->io_context->aic->nr_queued));
- atomic_dec(&arq->io_context->aic->nr_queued);
+ ioc = RQ_IOC(rq);
+ if (ioc && ioc->aic) {
+ BUG_ON(!atomic_read(&ioc->aic->nr_queued));
+ atomic_dec(&ioc->aic->nr_queued);
}
/*
- * Update the "next_arq" cache if we are about to remove its
+ * Update the "next_rq" cache if we are about to remove its
* entry
*/
- if (ad->next_arq[data_dir] == arq)
- ad->next_arq[data_dir] = as_find_next_arq(ad, arq);
+ if (ad->next_rq[data_dir] == rq)
+ ad->next_rq[data_dir] = as_find_next_rq(ad, rq);
rq_fifo_clear(rq);
- as_del_arq_rb(ad, rq);
+ as_del_rq_rb(ad, rq);
}
/*
- * as_fifo_expired returns 0 if there are no expired reads on the fifo,
+ * as_fifo_expired returns 0 if there are no expired requests on the fifo,
* 1 otherwise. It is ratelimited so that we only perform the check once per
* `fifo_expire' interval. Otherwise a large number of expired requests
* would create a hopeless seekstorm.
if (ad->changed_batch || ad->new_batch)
return 0;
- if (ad->batch_data_dir == REQ_SYNC)
+ if (ad->batch_data_dir == BLK_RW_SYNC)
/* TODO! add a check so a complete fifo gets written? */
return time_after(jiffies, ad->current_batch_expires);
/*
* move an entry to dispatch queue
*/
-static void as_move_to_dispatch(struct as_data *ad, struct as_rq *arq)
+static void as_move_to_dispatch(struct as_data *ad, struct request *rq)
{
- struct request *rq = arq->request;
const int data_dir = rq_is_sync(rq);
BUG_ON(RB_EMPTY_NODE(&rq->rb_node));
/*
* This has to be set in order to be correctly updated by
- * as_find_next_arq
+ * as_find_next_rq
*/
- ad->last_sector[data_dir] = rq->sector + rq->nr_sectors;
+ ad->last_sector[data_dir] = blk_rq_pos(rq) + blk_rq_sectors(rq);
- if (data_dir == REQ_SYNC) {
+ if (data_dir == BLK_RW_SYNC) {
+ struct io_context *ioc = RQ_IOC(rq);
/* In case we have to anticipate after this */
- copy_io_context(&ad->io_context, &arq->io_context);
+ copy_io_context(&ad->io_context, &ioc);
} else {
if (ad->io_context) {
put_io_context(ad->io_context);
}
ad->ioc_finished = 0;
- ad->next_arq[data_dir] = as_find_next_arq(ad, arq);
+ ad->next_rq[data_dir] = as_find_next_rq(ad, rq);
/*
* take it off the sort and fifo list, add to dispatch queue
*/
as_remove_queued_request(ad->q, rq);
- WARN_ON(arq->state != AS_RQ_QUEUED);
+ WARN_ON(RQ_STATE(rq) != AS_RQ_QUEUED);
elv_dispatch_sort(ad->q, rq);
- arq->state = AS_RQ_DISPATCHED;
- if (arq->io_context && arq->io_context->aic)
- atomic_inc(&arq->io_context->aic->nr_dispatched);
+ RQ_SET_STATE(rq, AS_RQ_DISPATCHED);
+ if (RQ_IOC(rq) && RQ_IOC(rq)->aic)
+ atomic_inc(&RQ_IOC(rq)->aic->nr_dispatched);
ad->nr_dispatched++;
}
* read/write expire, batch expire, etc, and moves it to the dispatch
* queue. Returns 1 if a request was found, 0 otherwise.
*/
-static int as_dispatch_request(request_queue_t *q, int force)
+static int as_dispatch_request(struct request_queue *q, int force)
{
struct as_data *ad = q->elevator->elevator_data;
- struct as_rq *arq;
- const int reads = !list_empty(&ad->fifo_list[REQ_SYNC]);
- const int writes = !list_empty(&ad->fifo_list[REQ_ASYNC]);
+ const int reads = !list_empty(&ad->fifo_list[BLK_RW_SYNC]);
+ const int writes = !list_empty(&ad->fifo_list[BLK_RW_ASYNC]);
+ struct request *rq;
if (unlikely(force)) {
/*
* Forced dispatch, accounting is useless. Reset
* accounting states and dump fifo_lists. Note that
- * batch_data_dir is reset to REQ_SYNC to avoid
+ * batch_data_dir is reset to BLK_RW_SYNC to avoid
* screwing write batch accounting as write batch
* accounting occurs on W->R transition.
*/
int dispatched = 0;
- ad->batch_data_dir = REQ_SYNC;
+ ad->batch_data_dir = BLK_RW_SYNC;
ad->changed_batch = 0;
ad->new_batch = 0;
- while (ad->next_arq[REQ_SYNC]) {
- as_move_to_dispatch(ad, ad->next_arq[REQ_SYNC]);
+ while (ad->next_rq[BLK_RW_SYNC]) {
+ as_move_to_dispatch(ad, ad->next_rq[BLK_RW_SYNC]);
dispatched++;
}
- ad->last_check_fifo[REQ_SYNC] = jiffies;
+ ad->last_check_fifo[BLK_RW_SYNC] = jiffies;
- while (ad->next_arq[REQ_ASYNC]) {
- as_move_to_dispatch(ad, ad->next_arq[REQ_ASYNC]);
+ while (ad->next_rq[BLK_RW_ASYNC]) {
+ as_move_to_dispatch(ad, ad->next_rq[BLK_RW_ASYNC]);
dispatched++;
}
- ad->last_check_fifo[REQ_ASYNC] = jiffies;
+ ad->last_check_fifo[BLK_RW_ASYNC] = jiffies;
return dispatched;
}
/* Signal that the write batch was uncontended, so we can't time it */
- if (ad->batch_data_dir == REQ_ASYNC && !reads) {
+ if (ad->batch_data_dir == BLK_RW_ASYNC && !reads) {
if (ad->current_write_count == 0 || !writes)
ad->write_batch_idled = 1;
}
/*
* batch is still running or no reads or no writes
*/
- arq = ad->next_arq[ad->batch_data_dir];
+ rq = ad->next_rq[ad->batch_data_dir];
- if (ad->batch_data_dir == REQ_SYNC && ad->antic_expire) {
- if (as_fifo_expired(ad, REQ_SYNC))
+ if (ad->batch_data_dir == BLK_RW_SYNC && ad->antic_expire) {
+ if (as_fifo_expired(ad, BLK_RW_SYNC))
goto fifo_expired;
- if (as_can_anticipate(ad, arq)) {
+ if (as_can_anticipate(ad, rq)) {
as_antic_waitreq(ad);
return 0;
}
}
- if (arq) {
+ if (rq) {
/* we have a "next request" */
if (reads && !writes)
ad->current_batch_expires =
- jiffies + ad->batch_expire[REQ_SYNC];
+ jiffies + ad->batch_expire[BLK_RW_SYNC];
goto dispatch_request;
}
}
*/
if (reads) {
- BUG_ON(RB_EMPTY_ROOT(&ad->sort_list[REQ_SYNC]));
+ BUG_ON(RB_EMPTY_ROOT(&ad->sort_list[BLK_RW_SYNC]));
- if (writes && ad->batch_data_dir == REQ_SYNC)
+ if (writes && ad->batch_data_dir == BLK_RW_SYNC)
/*
* Last batch was a read, switch to writes
*/
goto dispatch_writes;
- if (ad->batch_data_dir == REQ_ASYNC) {
+ if (ad->batch_data_dir == BLK_RW_ASYNC) {
WARN_ON(ad->new_batch);
ad->changed_batch = 1;
}
- ad->batch_data_dir = REQ_SYNC;
- arq = RQ_DATA(rq_entry_fifo(ad->fifo_list[REQ_SYNC].next));
+ ad->batch_data_dir = BLK_RW_SYNC;
+ rq = rq_entry_fifo(ad->fifo_list[BLK_RW_SYNC].next);
ad->last_check_fifo[ad->batch_data_dir] = jiffies;
goto dispatch_request;
}
if (writes) {
dispatch_writes:
- BUG_ON(RB_EMPTY_ROOT(&ad->sort_list[REQ_ASYNC]));
+ BUG_ON(RB_EMPTY_ROOT(&ad->sort_list[BLK_RW_ASYNC]));
- if (ad->batch_data_dir == REQ_SYNC) {
+ if (ad->batch_data_dir == BLK_RW_SYNC) {
ad->changed_batch = 1;
/*
*/
ad->new_batch = 0;
}
- ad->batch_data_dir = REQ_ASYNC;
+ ad->batch_data_dir = BLK_RW_ASYNC;
ad->current_write_count = ad->write_batch_count;
ad->write_batch_idled = 0;
- arq = ad->next_arq[ad->batch_data_dir];
+ rq = rq_entry_fifo(ad->fifo_list[BLK_RW_ASYNC].next);
+ ad->last_check_fifo[BLK_RW_ASYNC] = jiffies;
goto dispatch_request;
}
if (as_fifo_expired(ad, ad->batch_data_dir)) {
fifo_expired:
- arq = RQ_DATA(rq_entry_fifo(ad->fifo_list[ad->batch_data_dir].next));
+ rq = rq_entry_fifo(ad->fifo_list[ad->batch_data_dir].next);
}
if (ad->changed_batch) {
if (ad->nr_dispatched)
return 0;
- if (ad->batch_data_dir == REQ_ASYNC)
+ if (ad->batch_data_dir == BLK_RW_ASYNC)
ad->current_batch_expires = jiffies +
- ad->batch_expire[REQ_ASYNC];
+ ad->batch_expire[BLK_RW_ASYNC];
else
ad->new_batch = 1;
}
/*
- * arq is the selected appropriate request.
+ * rq is the selected appropriate request.
*/
- as_move_to_dispatch(ad, arq);
+ as_move_to_dispatch(ad, rq);
return 1;
}
/*
- * add arq to rbtree and fifo
+ * add rq to rbtree and fifo
*/
-static void as_add_request(request_queue_t *q, struct request *rq)
+static void as_add_request(struct request_queue *q, struct request *rq)
{
struct as_data *ad = q->elevator->elevator_data;
- struct as_rq *arq = RQ_DATA(rq);
int data_dir;
- arq->state = AS_RQ_NEW;
+ RQ_SET_STATE(rq, AS_RQ_NEW);
data_dir = rq_is_sync(rq);
- arq->io_context = as_get_io_context();
+ rq->elevator_private = as_get_io_context(q->node);
- if (arq->io_context) {
- as_update_iohist(ad, arq->io_context->aic, arq->request);
- atomic_inc(&arq->io_context->aic->nr_queued);
+ if (RQ_IOC(rq)) {
+ as_update_iohist(ad, RQ_IOC(rq)->aic, rq);
+ atomic_inc(&RQ_IOC(rq)->aic->nr_queued);
}
- as_add_arq_rb(ad, rq);
+ as_add_rq_rb(ad, rq);
/*
- * set expire time (only used for reads) and add to fifo list
+ * set expire time and add to fifo list
*/
rq_set_fifo_time(rq, jiffies + ad->fifo_expire[data_dir]);
list_add_tail(&rq->queuelist, &ad->fifo_list[data_dir]);
- as_update_arq(ad, arq); /* keep state machine up to date */
- arq->state = AS_RQ_QUEUED;
+ as_update_rq(ad, rq); /* keep state machine up to date */
+ RQ_SET_STATE(rq, AS_RQ_QUEUED);
}
-static void as_activate_request(request_queue_t *q, struct request *rq)
+static void as_activate_request(struct request_queue *q, struct request *rq)
{
- struct as_rq *arq = RQ_DATA(rq);
-
- WARN_ON(arq->state != AS_RQ_DISPATCHED);
- arq->state = AS_RQ_REMOVED;
- if (arq->io_context && arq->io_context->aic)
- atomic_dec(&arq->io_context->aic->nr_dispatched);
+ WARN_ON(RQ_STATE(rq) != AS_RQ_DISPATCHED);
+ RQ_SET_STATE(rq, AS_RQ_REMOVED);
+ if (RQ_IOC(rq) && RQ_IOC(rq)->aic)
+ atomic_dec(&RQ_IOC(rq)->aic->nr_dispatched);
}
-static void as_deactivate_request(request_queue_t *q, struct request *rq)
+static void as_deactivate_request(struct request_queue *q, struct request *rq)
{
- struct as_rq *arq = RQ_DATA(rq);
-
- WARN_ON(arq->state != AS_RQ_REMOVED);
- arq->state = AS_RQ_DISPATCHED;
- if (arq->io_context && arq->io_context->aic)
- atomic_inc(&arq->io_context->aic->nr_dispatched);
+ WARN_ON(RQ_STATE(rq) != AS_RQ_REMOVED);
+ RQ_SET_STATE(rq, AS_RQ_DISPATCHED);
+ if (RQ_IOC(rq) && RQ_IOC(rq)->aic)
+ atomic_inc(&RQ_IOC(rq)->aic->nr_dispatched);
}
/*
* is not empty - it is used in the block layer to check for plugging and
* merging opportunities
*/
-static int as_queue_empty(request_queue_t *q)
+static int as_queue_empty(struct request_queue *q)
{
struct as_data *ad = q->elevator->elevator_data;
- return list_empty(&ad->fifo_list[REQ_ASYNC])
- && list_empty(&ad->fifo_list[REQ_SYNC]);
+ return list_empty(&ad->fifo_list[BLK_RW_ASYNC])
+ && list_empty(&ad->fifo_list[BLK_RW_SYNC]);
}
static int
-as_merge(request_queue_t *q, struct request **req, struct bio *bio)
+as_merge(struct request_queue *q, struct request **req, struct bio *bio)
{
struct as_data *ad = q->elevator->elevator_data;
sector_t rb_key = bio->bi_sector + bio_sectors(bio);
return ELEVATOR_NO_MERGE;
}
-static void as_merged_request(request_queue_t *q, struct request *req, int type)
+static void as_merged_request(struct request_queue *q, struct request *req,
+ int type)
{
struct as_data *ad = q->elevator->elevator_data;
* if the merge was a front merge, we need to reposition request
*/
if (type == ELEVATOR_FRONT_MERGE) {
- as_del_arq_rb(ad, req);
- as_add_arq_rb(ad, req);
+ as_del_rq_rb(ad, req);
+ as_add_rq_rb(ad, req);
/*
* Note! At this stage of this and the next function, our next
* request may not be optimal - eg the request may have "grown"
}
}
-static void as_merged_requests(request_queue_t *q, struct request *req,
+static void as_merged_requests(struct request_queue *q, struct request *req,
struct request *next)
{
- struct as_rq *arq = RQ_DATA(req);
- struct as_rq *anext = RQ_DATA(next);
-
- BUG_ON(!arq);
- BUG_ON(!anext);
-
/*
- * if anext expires before arq, assign its expire time to arq
- * and move into anext position (anext will be deleted) in fifo
+ * if next expires before rq, assign its expire time to arq
+ * and move into next position (next will be deleted) in fifo
*/
if (!list_empty(&req->queuelist) && !list_empty(&next->queuelist)) {
if (time_before(rq_fifo_time(next), rq_fifo_time(req))) {
list_move(&req->queuelist, &next->queuelist);
rq_set_fifo_time(req, rq_fifo_time(next));
- /*
- * Don't copy here but swap, because when anext is
- * removed below, it must contain the unused context
- */
- swap_io_context(&arq->io_context, &anext->io_context);
}
}
* kill knowledge of next, this one is a goner
*/
as_remove_queued_request(q, next);
- as_put_io_context(anext);
+ as_put_io_context(next);
- anext->state = AS_RQ_MERGED;
+ RQ_SET_STATE(next, AS_RQ_MERGED);
}
/*
*
* FIXME! dispatch queue is not a queue at all!
*/
-static void as_work_handler(void *data)
-{
- struct request_queue *q = data;
- unsigned long flags;
-
- spin_lock_irqsave(q->queue_lock, flags);
- if (!as_queue_empty(q))
- q->request_fn(q);
- spin_unlock_irqrestore(q->queue_lock, flags);
-}
-
-static void as_put_request(request_queue_t *q, struct request *rq)
+static void as_work_handler(struct work_struct *work)
{
- struct as_data *ad = q->elevator->elevator_data;
- struct as_rq *arq = RQ_DATA(rq);
+ struct as_data *ad = container_of(work, struct as_data, antic_work);
- if (!arq) {
- WARN_ON(1);
- return;
- }
-
- if (unlikely(arq->state != AS_RQ_POSTSCHED &&
- arq->state != AS_RQ_PRESCHED &&
- arq->state != AS_RQ_MERGED)) {
- printk("arq->state %d\n", arq->state);
- WARN_ON(1);
- }
-
- mempool_free(arq, ad->arq_pool);
- rq->elevator_private = NULL;
+ blk_run_queue(ad->q);
}
-static int as_set_request(request_queue_t *q, struct request *rq,
- struct bio *bio, gfp_t gfp_mask)
-{
- struct as_data *ad = q->elevator->elevator_data;
- struct as_rq *arq = mempool_alloc(ad->arq_pool, gfp_mask);
-
- if (arq) {
- memset(arq, 0, sizeof(*arq));
- arq->request = rq;
- arq->state = AS_RQ_PRESCHED;
- arq->io_context = NULL;
- rq->elevator_private = arq;
- return 0;
- }
-
- return 1;
-}
-
-static int as_may_queue(request_queue_t *q, int rw, struct bio *bio)
+static int as_may_queue(struct request_queue *q, int rw)
{
int ret = ELV_MQUEUE_MAY;
struct as_data *ad = q->elevator->elevator_data;
struct io_context *ioc;
if (ad->antic_status == ANTIC_WAIT_REQ ||
ad->antic_status == ANTIC_WAIT_NEXT) {
- ioc = as_get_io_context();
+ ioc = as_get_io_context(q->node);
if (ad->io_context == ioc)
ret = ELV_MQUEUE_MUST;
put_io_context(ioc);
return ret;
}
-static void as_exit_queue(elevator_t *e)
+static void as_exit_queue(struct elevator_queue *e)
{
struct as_data *ad = e->elevator_data;
del_timer_sync(&ad->antic_timer);
- kblockd_flush();
+ cancel_work_sync(&ad->antic_work);
- BUG_ON(!list_empty(&ad->fifo_list[REQ_SYNC]));
- BUG_ON(!list_empty(&ad->fifo_list[REQ_ASYNC]));
+ BUG_ON(!list_empty(&ad->fifo_list[BLK_RW_SYNC]));
+ BUG_ON(!list_empty(&ad->fifo_list[BLK_RW_ASYNC]));
- mempool_destroy(ad->arq_pool);
put_io_context(ad->io_context);
kfree(ad);
}
/*
- * initialize elevator private data (as_data), and alloc a arq for
- * each request on the free lists
+ * initialize elevator private data (as_data).
*/
-static void *as_init_queue(request_queue_t *q, elevator_t *e)
+static void *as_init_queue(struct request_queue *q)
{
struct as_data *ad;
- if (!arq_pool)
- return NULL;
-
- ad = kmalloc_node(sizeof(*ad), GFP_KERNEL, q->node);
+ ad = kmalloc_node(sizeof(*ad), GFP_KERNEL | __GFP_ZERO, q->node);
if (!ad)
return NULL;
- memset(ad, 0, sizeof(*ad));
ad->q = q; /* Identify what queue the data belongs to */
- ad->arq_pool = mempool_create_node(BLKDEV_MIN_RQ, mempool_alloc_slab,
- mempool_free_slab, arq_pool, q->node);
- if (!ad->arq_pool) {
- kfree(ad);
- return NULL;
- }
-
/* anticipatory scheduling helpers */
ad->antic_timer.function = as_antic_timeout;
ad->antic_timer.data = (unsigned long)q;
init_timer(&ad->antic_timer);
- INIT_WORK(&ad->antic_work, as_work_handler, q);
-
- INIT_LIST_HEAD(&ad->fifo_list[REQ_SYNC]);
- INIT_LIST_HEAD(&ad->fifo_list[REQ_ASYNC]);
- ad->sort_list[REQ_SYNC] = RB_ROOT;
- ad->sort_list[REQ_ASYNC] = RB_ROOT;
- ad->fifo_expire[REQ_SYNC] = default_read_expire;
- ad->fifo_expire[REQ_ASYNC] = default_write_expire;
+ INIT_WORK(&ad->antic_work, as_work_handler);
+
+ INIT_LIST_HEAD(&ad->fifo_list[BLK_RW_SYNC]);
+ INIT_LIST_HEAD(&ad->fifo_list[BLK_RW_ASYNC]);
+ ad->sort_list[BLK_RW_SYNC] = RB_ROOT;
+ ad->sort_list[BLK_RW_ASYNC] = RB_ROOT;
+ ad->fifo_expire[BLK_RW_SYNC] = default_read_expire;
+ ad->fifo_expire[BLK_RW_ASYNC] = default_write_expire;
ad->antic_expire = default_antic_expire;
- ad->batch_expire[REQ_SYNC] = default_read_batch_expire;
- ad->batch_expire[REQ_ASYNC] = default_write_batch_expire;
+ ad->batch_expire[BLK_RW_SYNC] = default_read_batch_expire;
+ ad->batch_expire[BLK_RW_ASYNC] = default_write_batch_expire;
- ad->current_batch_expires = jiffies + ad->batch_expire[REQ_SYNC];
- ad->write_batch_count = ad->batch_expire[REQ_ASYNC] / 10;
+ ad->current_batch_expires = jiffies + ad->batch_expire[BLK_RW_SYNC];
+ ad->write_batch_count = ad->batch_expire[BLK_RW_ASYNC] / 10;
if (ad->write_batch_count < 2)
ad->write_batch_count = 2;
return count;
}
-static ssize_t est_time_show(elevator_t *e, char *page)
+static ssize_t est_time_show(struct elevator_queue *e, char *page)
{
struct as_data *ad = e->elevator_data;
int pos = 0;
}
#define SHOW_FUNCTION(__FUNC, __VAR) \
-static ssize_t __FUNC(elevator_t *e, char *page) \
+static ssize_t __FUNC(struct elevator_queue *e, char *page) \
{ \
struct as_data *ad = e->elevator_data; \
return as_var_show(jiffies_to_msecs((__VAR)), (page)); \
}
-SHOW_FUNCTION(as_read_expire_show, ad->fifo_expire[REQ_SYNC]);
-SHOW_FUNCTION(as_write_expire_show, ad->fifo_expire[REQ_ASYNC]);
+SHOW_FUNCTION(as_read_expire_show, ad->fifo_expire[BLK_RW_SYNC]);
+SHOW_FUNCTION(as_write_expire_show, ad->fifo_expire[BLK_RW_ASYNC]);
SHOW_FUNCTION(as_antic_expire_show, ad->antic_expire);
-SHOW_FUNCTION(as_read_batch_expire_show, ad->batch_expire[REQ_SYNC]);
-SHOW_FUNCTION(as_write_batch_expire_show, ad->batch_expire[REQ_ASYNC]);
+SHOW_FUNCTION(as_read_batch_expire_show, ad->batch_expire[BLK_RW_SYNC]);
+SHOW_FUNCTION(as_write_batch_expire_show, ad->batch_expire[BLK_RW_ASYNC]);
#undef SHOW_FUNCTION
#define STORE_FUNCTION(__FUNC, __PTR, MIN, MAX) \
-static ssize_t __FUNC(elevator_t *e, const char *page, size_t count) \
+static ssize_t __FUNC(struct elevator_queue *e, const char *page, size_t count) \
{ \
struct as_data *ad = e->elevator_data; \
int ret = as_var_store(__PTR, (page), count); \
*(__PTR) = msecs_to_jiffies(*(__PTR)); \
return ret; \
}
-STORE_FUNCTION(as_read_expire_store, &ad->fifo_expire[REQ_SYNC], 0, INT_MAX);
-STORE_FUNCTION(as_write_expire_store, &ad->fifo_expire[REQ_ASYNC], 0, INT_MAX);
+STORE_FUNCTION(as_read_expire_store, &ad->fifo_expire[BLK_RW_SYNC], 0, INT_MAX);
+STORE_FUNCTION(as_write_expire_store,
+ &ad->fifo_expire[BLK_RW_ASYNC], 0, INT_MAX);
STORE_FUNCTION(as_antic_expire_store, &ad->antic_expire, 0, INT_MAX);
STORE_FUNCTION(as_read_batch_expire_store,
- &ad->batch_expire[REQ_SYNC], 0, INT_MAX);
+ &ad->batch_expire[BLK_RW_SYNC], 0, INT_MAX);
STORE_FUNCTION(as_write_batch_expire_store,
- &ad->batch_expire[REQ_ASYNC], 0, INT_MAX);
+ &ad->batch_expire[BLK_RW_ASYNC], 0, INT_MAX);
#undef STORE_FUNCTION
#define AS_ATTR(name) \
.elevator_completed_req_fn = as_completed_request,
.elevator_former_req_fn = elv_rb_former_request,
.elevator_latter_req_fn = elv_rb_latter_request,
- .elevator_set_req_fn = as_set_request,
- .elevator_put_req_fn = as_put_request,
.elevator_may_queue_fn = as_may_queue,
.elevator_init_fn = as_init_queue,
.elevator_exit_fn = as_exit_queue,
static int __init as_init(void)
{
- int ret;
+ elv_register(&iosched_as);
- arq_pool = kmem_cache_create("as_arq", sizeof(struct as_rq),
- 0, 0, NULL, NULL);
- if (!arq_pool)
- return -ENOMEM;
-
- ret = elv_register(&iosched_as);
- if (!ret) {
- /*
- * don't allow AS to get unregistered, since we would have
- * to browse all tasks in the system and release their
- * as_io_context first
- */
- __module_get(THIS_MODULE);
- return 0;
- }
-
- kmem_cache_destroy(arq_pool);
- return ret;
+ return 0;
}
static void __exit as_exit(void)
{
- DECLARE_COMPLETION(all_gone);
+ DECLARE_COMPLETION_ONSTACK(all_gone);
elv_unregister(&iosched_as);
ioc_gone = &all_gone;
/* ioc_gone's update must be visible before reading ioc_count */
smp_wmb();
- if (atomic_read(&ioc_count))
- wait_for_completion(ioc_gone);
+ if (elv_ioc_count_read(as_ioc_count))
+ wait_for_completion(&all_gone);
synchronize_rcu();
- kmem_cache_destroy(arq_pool);
}
module_init(as_init);