/*
- * linux/drivers/block/as-iosched.c
- *
* Anticipatory & deadline i/o scheduler.
*
* Copyright (C) 2002 Jens Axboe <axboe@suse.de>
- * Nick Piggin <piggin@cyberone.com.au>
+ * Nick Piggin <nickpiggin@yahoo.com.au>
*
*/
#include <linux/kernel.h>
#include <linux/blkdev.h>
#include <linux/elevator.h>
#include <linux/bio.h>
-#include <linux/config.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/compiler.h>
-#include <linux/hash.h>
#include <linux/rbtree.h>
#include <linux/interrupt.h>
/* Bits in as_io_context.state */
enum as_io_states {
- AS_TASK_RUNNING=0, /* Process has not exitted */
+ AS_TASK_RUNNING=0, /* Process has not exited */
AS_TASK_IOSTARTED, /* Process has started some IO */
AS_TASK_IORUNNING, /* Process has completed some IO */
};
struct as_rq *next_arq[2]; /* next in sort order */
sector_t last_sector[2]; /* last REQ_SYNC & REQ_ASYNC sectors */
- struct list_head *hash; /* request hash */
unsigned long exit_prob; /* probability a task will exit while
being waited on */
+ unsigned long exit_no_coop; /* probablility an exited task will
+ not be part of a later cooperating
+ request */
unsigned long new_ttime_total; /* mean thinktime on new proc */
unsigned long new_ttime_mean;
u64 new_seek_total; /* mean seek on new proc */
};
struct as_rq {
- /*
- * rbtree index, key is the starting offset
- */
- struct rb_node rb_node;
- sector_t rb_key;
-
struct request *request;
struct io_context *io_context; /* The submitting task */
/*
- * request hash, key is the ending offset (for back merge lookup)
- */
- struct list_head hash;
- unsigned int on_hash;
-
- /*
* expire fifo
*/
struct list_head fifo;
static kmem_cache_t *arq_pool;
+static atomic_t ioc_count = ATOMIC_INIT(0);
+static struct completion *ioc_gone;
+
+static void as_move_to_dispatch(struct as_data *ad, struct as_rq *arq);
+static void as_antic_stop(struct as_data *ad);
+
/*
* IO Context helper functions
*/
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);
+}
+
+static void as_trim(struct io_context *ioc)
+{
+ if (ioc->aic)
+ free_as_io_context(ioc->aic);
+ ioc->aic = NULL;
}
/* Called when the task exits */
ret->seek_total = 0;
ret->seek_samples = 0;
ret->seek_mean = 0;
+ atomic_inc(&ioc_count);
}
return ret;
}
/*
- * the back merge hash support functions
- */
-static const int as_hash_shift = 6;
-#define AS_HASH_BLOCK(sec) ((sec) >> 3)
-#define AS_HASH_FN(sec) (hash_long(AS_HASH_BLOCK((sec)), as_hash_shift))
-#define AS_HASH_ENTRIES (1 << as_hash_shift)
-#define rq_hash_key(rq) ((rq)->sector + (rq)->nr_sectors)
-#define list_entry_hash(ptr) list_entry((ptr), struct as_rq, hash)
-
-static inline void __as_del_arq_hash(struct as_rq *arq)
-{
- arq->on_hash = 0;
- list_del_init(&arq->hash);
-}
-
-static inline void as_del_arq_hash(struct as_rq *arq)
-{
- if (arq->on_hash)
- __as_del_arq_hash(arq);
-}
-
-static void as_add_arq_hash(struct as_data *ad, struct as_rq *arq)
-{
- struct request *rq = arq->request;
-
- BUG_ON(arq->on_hash);
-
- arq->on_hash = 1;
- list_add(&arq->hash, &ad->hash[AS_HASH_FN(rq_hash_key(rq))]);
-}
-
-/*
- * move hot entry to front of chain
- */
-static inline void as_hot_arq_hash(struct as_data *ad, struct as_rq *arq)
-{
- struct request *rq = arq->request;
- struct list_head *head = &ad->hash[AS_HASH_FN(rq_hash_key(rq))];
-
- if (!arq->on_hash) {
- WARN_ON(1);
- return;
- }
-
- if (arq->hash.prev != head) {
- list_del(&arq->hash);
- list_add(&arq->hash, head);
- }
-}
-
-static struct request *as_find_arq_hash(struct as_data *ad, sector_t offset)
-{
- struct list_head *hash_list = &ad->hash[AS_HASH_FN(offset)];
- struct list_head *entry, *next = hash_list->next;
-
- while ((entry = next) != hash_list) {
- struct as_rq *arq = list_entry_hash(entry);
- struct request *__rq = arq->request;
-
- next = entry->next;
-
- BUG_ON(!arq->on_hash);
-
- if (!rq_mergeable(__rq)) {
- as_del_arq_hash(arq);
- continue;
- }
-
- if (rq_hash_key(__rq) == offset)
- return __rq;
- }
-
- return NULL;
-}
-
-/*
* rb tree support functions
*/
-#define RB_NONE (2)
-#define RB_EMPTY(root) ((root)->rb_node == NULL)
-#define ON_RB(node) ((node)->rb_color != RB_NONE)
-#define RB_CLEAR(node) ((node)->rb_color = RB_NONE)
-#define rb_entry_arq(node) rb_entry((node), struct as_rq, rb_node)
#define ARQ_RB_ROOT(ad, arq) (&(ad)->sort_list[(arq)->is_sync])
-#define rq_rb_key(rq) (rq)->sector
-/*
- * as_find_first_arq finds the first (lowest sector numbered) request
- * for the specified data_dir. Used to sweep back to the start of the disk
- * (1-way elevator) after we process the last (highest sector) request.
- */
-static struct as_rq *as_find_first_arq(struct as_data *ad, int data_dir)
+static void as_add_arq_rb(struct as_data *ad, struct request *rq)
{
- struct rb_node *n = ad->sort_list[data_dir].rb_node;
-
- if (n == NULL)
- return NULL;
-
- for (;;) {
- if (n->rb_left == NULL)
- return rb_entry_arq(n);
-
- n = n->rb_left;
- }
-}
-
-/*
- * Add the request to the rb tree if it is unique. If there is an alias (an
- * existing request against the same sector), which can happen when using
- * direct IO, then return the alias.
- */
-static struct as_rq *as_add_arq_rb(struct as_data *ad, struct as_rq *arq)
-{
- struct rb_node **p = &ARQ_RB_ROOT(ad, arq)->rb_node;
- struct rb_node *parent = NULL;
- struct as_rq *__arq;
- struct request *rq = arq->request;
-
- arq->rb_key = rq_rb_key(rq);
-
- while (*p) {
- parent = *p;
- __arq = rb_entry_arq(parent);
-
- if (arq->rb_key < __arq->rb_key)
- p = &(*p)->rb_left;
- else if (arq->rb_key > __arq->rb_key)
- p = &(*p)->rb_right;
- else
- return __arq;
- }
-
- rb_link_node(&arq->rb_node, parent, p);
- rb_insert_color(&arq->rb_node, ARQ_RB_ROOT(ad, arq));
-
- return NULL;
-}
+ struct as_rq *arq = RQ_DATA(rq);
+ struct request *alias;
-static inline void as_del_arq_rb(struct as_data *ad, struct as_rq *arq)
-{
- if (!ON_RB(&arq->rb_node)) {
- WARN_ON(1);
- return;
+ while ((unlikely(alias = elv_rb_add(ARQ_RB_ROOT(ad, arq), rq)))) {
+ as_move_to_dispatch(ad, RQ_DATA(alias));
+ as_antic_stop(ad);
}
-
- rb_erase(&arq->rb_node, ARQ_RB_ROOT(ad, arq));
- RB_CLEAR(&arq->rb_node);
}
-static struct request *
-as_find_arq_rb(struct as_data *ad, sector_t sector, int data_dir)
+static inline void as_del_arq_rb(struct as_data *ad, struct request *rq)
{
- struct rb_node *n = ad->sort_list[data_dir].rb_node;
- struct as_rq *arq;
-
- while (n) {
- arq = rb_entry_arq(n);
-
- if (sector < arq->rb_key)
- n = n->rb_left;
- else if (sector > arq->rb_key)
- n = n->rb_right;
- else
- return arq->request;
- }
-
- return NULL;
+ elv_rb_del(ARQ_RB_ROOT(ad, RQ_DATA(rq)), rq);
}
/*
* 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 *last)
+static struct as_rq *as_find_next_arq(struct as_data *ad, struct as_rq *arq)
{
- const int data_dir = last->is_sync;
- struct as_rq *ret;
+ 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 *arq_next, *arq_prev;
+ struct as_rq *next = NULL, *prev = NULL;
- BUG_ON(!ON_RB(&last->rb_node));
+ BUG_ON(RB_EMPTY_NODE(&last->rb_node));
if (rbprev)
- arq_prev = rb_entry_arq(rbprev);
- else
- arq_prev = NULL;
+ prev = RQ_DATA(rb_entry_rq(rbprev));
if (rbnext)
- arq_next = rb_entry_arq(rbnext);
+ next = RQ_DATA(rb_entry_rq(rbnext));
else {
- arq_next = as_find_first_arq(ad, data_dir);
- if (arq_next == last)
- arq_next = NULL;
- }
+ const int data_dir = arq->is_sync;
- ret = as_choose_req(ad, arq_next, arq_prev);
+ rbnext = rb_first(&ad->sort_list[data_dir]);
+ if (rbnext && rbnext != &last->rb_node)
+ next = RQ_DATA(rb_entry_rq(rbnext));
+ }
- return ret;
+ return as_choose_req(ad, next, prev);
}
/*
kblockd_schedule_work(&ad->antic_work);
if (aic->ttime_samples == 0) {
- /* process anticipated on has exitted or timed out*/
+ /* process anticipated on has exited or timed out*/
ad->exit_prob = (7*ad->exit_prob + 256)/8;
}
+ if (!test_bit(AS_TASK_RUNNING, &aic->state)) {
+ /* process not "saved" by a cooperating request */
+ ad->exit_no_coop = (7*ad->exit_no_coop + 256)/8;
+ }
}
spin_unlock_irqrestore(q->queue_lock, flags);
}
+static void as_update_thinktime(struct as_data *ad, struct as_io_context *aic,
+ unsigned long ttime)
+{
+ /* fixed point: 1.0 == 1<<8 */
+ if (aic->ttime_samples == 0) {
+ ad->new_ttime_total = (7*ad->new_ttime_total + 256*ttime) / 8;
+ ad->new_ttime_mean = ad->new_ttime_total / 256;
+
+ ad->exit_prob = (7*ad->exit_prob)/8;
+ }
+ aic->ttime_samples = (7*aic->ttime_samples + 256) / 8;
+ aic->ttime_total = (7*aic->ttime_total + 256*ttime) / 8;
+ aic->ttime_mean = (aic->ttime_total + 128) / aic->ttime_samples;
+}
+
+static void as_update_seekdist(struct as_data *ad, struct as_io_context *aic,
+ sector_t sdist)
+{
+ u64 total;
+
+ if (aic->seek_samples == 0) {
+ ad->new_seek_total = (7*ad->new_seek_total + 256*(u64)sdist)/8;
+ ad->new_seek_mean = ad->new_seek_total / 256;
+ }
+
+ /*
+ * Don't allow the seek distance to get too large from the
+ * odd fragment, pagein, etc
+ */
+ if (aic->seek_samples <= 60) /* second&third seek */
+ sdist = min(sdist, (aic->seek_mean * 4) + 2*1024*1024);
+ else
+ sdist = min(sdist, (aic->seek_mean * 4) + 2*1024*64);
+
+ aic->seek_samples = (7*aic->seek_samples + 256) / 8;
+ aic->seek_total = (7*aic->seek_total + (u64)256*sdist) / 8;
+ total = aic->seek_total + (aic->seek_samples/2);
+ do_div(total, aic->seek_samples);
+ aic->seek_mean = (sector_t)total;
+}
+
+/*
+ * as_update_iohist keeps a decaying histogram of IO thinktimes, and
+ * updates @aic->ttime_mean based on that. It is called when a new
+ * request is queued.
+ */
+static void as_update_iohist(struct as_data *ad, struct as_io_context *aic,
+ struct request *rq)
+{
+ struct as_rq *arq = RQ_DATA(rq);
+ int data_dir = arq->is_sync;
+ unsigned long thinktime = 0;
+ sector_t seek_dist;
+
+ if (aic == NULL)
+ return;
+
+ if (data_dir == REQ_SYNC) {
+ unsigned long in_flight = atomic_read(&aic->nr_queued)
+ + atomic_read(&aic->nr_dispatched);
+ spin_lock(&aic->lock);
+ if (test_bit(AS_TASK_IORUNNING, &aic->state) ||
+ test_bit(AS_TASK_IOSTARTED, &aic->state)) {
+ /* Calculate read -> read thinktime */
+ if (test_bit(AS_TASK_IORUNNING, &aic->state)
+ && in_flight == 0) {
+ thinktime = jiffies - aic->last_end_request;
+ thinktime = min(thinktime, MAX_THINKTIME-1);
+ }
+ 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;
+ else
+ seek_dist = aic->last_request_pos - rq->sector;
+ as_update_seekdist(ad, aic, seek_dist);
+ }
+ aic->last_request_pos = rq->sector + rq->nr_sectors;
+ set_bit(AS_TASK_IOSTARTED, &aic->state);
+ spin_unlock(&aic->lock);
+ }
+}
+
/*
* as_close_req decides if one request is considered "close" to the
* previous one issued.
*/
-static int as_close_req(struct as_data *ad, struct as_rq *arq)
+static int as_close_req(struct as_data *ad, struct as_io_context *aic,
+ struct as_rq *arq)
{
unsigned long delay; /* milliseconds */
sector_t last = ad->last_sector[ad->batch_data_dir];
sector_t next = arq->request->sector;
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;
- if (delay <= 1)
- delta = 64;
+ if (delay == 0)
+ delta = 8192;
else if (delay <= 20 && delay <= ad->antic_expire)
- delta = 64 << (delay-1);
+ delta = 8192 << delay;
else
return 1;
- return (last - (delta>>1) <= next) && (next <= last + delta);
+ if ((last <= next + (delta>>1)) && (next <= last + delta))
+ return 1;
+
+ if (last < next)
+ s = next - last;
+ else
+ s = last - next;
+
+ if (aic->seek_samples == 0) {
+ /*
+ * Process has just started IO. Use past statistics to
+ * gauge success possibility
+ */
+ if (ad->new_seek_mean > s) {
+ /* this request is better than what we're expecting */
+ return 1;
+ }
+
+ } else {
+ if (aic->seek_mean > s) {
+ /* this request is better than what we're expecting */
+ return 1;
+ }
+ }
+
+ return 0;
}
/*
* dispatch it ASAP, because we know that application will not be submitting
* any new reads.
*
- * If the task which has submitted the request has exitted, break anticipation.
+ * If the task which has submitted the request has exited, break anticipation.
*
* If this task has queued some other IO, do not enter enticipation.
*/
{
struct io_context *ioc;
struct as_io_context *aic;
- sector_t s;
ioc = ad->io_context;
BUG_ON(!ioc);
if (!aic)
return 0;
- if (!test_bit(AS_TASK_RUNNING, &aic->state)) {
- /* process anticipated on has exitted */
- if (aic->ttime_samples == 0)
- ad->exit_prob = (7*ad->exit_prob + 256)/8;
- return 1;
- }
-
if (atomic_read(&aic->nr_queued) > 0) {
/* process has more requests queued */
return 1;
return 1;
}
- if (arq && arq->is_sync == REQ_SYNC && as_close_req(ad, arq)) {
+ if (arq && arq->is_sync == REQ_SYNC && as_close_req(ad, aic, arq)) {
/*
* Found a close request that is not one of ours.
*
- * This makes close requests from another process reset
- * our thinktime delay. Is generally useful when there are
+ * This makes close requests from another process update
+ * our IO history. Is generally useful when there are
* two or more cooperating processes working in the same
* area.
*/
- spin_lock(&aic->lock);
- aic->last_end_request = jiffies;
- spin_unlock(&aic->lock);
+ if (!test_bit(AS_TASK_RUNNING, &aic->state)) {
+ if (aic->ttime_samples == 0)
+ ad->exit_prob = (7*ad->exit_prob + 256)/8;
+
+ ad->exit_no_coop = (7*ad->exit_no_coop)/8;
+ }
+
+ as_update_iohist(ad, aic, arq->request);
return 1;
}
+ if (!test_bit(AS_TASK_RUNNING, &aic->state)) {
+ /* process anticipated on has exited */
+ if (aic->ttime_samples == 0)
+ ad->exit_prob = (7*ad->exit_prob + 256)/8;
+
+ if (ad->exit_no_coop > 128)
+ return 1;
+ }
if (aic->ttime_samples == 0) {
if (ad->new_ttime_mean > ad->antic_expire)
return 1;
- if (ad->exit_prob > 128)
+ if (ad->exit_prob * ad->exit_no_coop > 128*256)
return 1;
} else if (aic->ttime_mean > ad->antic_expire) {
/* the process thinks too much between requests */
return 1;
}
- if (!arq)
- return 0;
-
- if (ad->last_sector[REQ_SYNC] < arq->request->sector)
- s = arq->request->sector - ad->last_sector[REQ_SYNC];
- else
- s = ad->last_sector[REQ_SYNC] - arq->request->sector;
-
- if (aic->seek_samples == 0) {
- /*
- * Process has just started IO. Use past statistics to
- * guage success possibility
- */
- if (ad->new_seek_mean > s) {
- /* this request is better than what we're expecting */
- return 1;
- }
-
- } else {
- if (aic->seek_mean > s) {
- /* this request is better than what we're expecting */
- return 1;
- }
- }
-
return 0;
}
/*
- * as_can_anticipate indicates weather we should either run arq
+ * as_can_anticipate indicates whether we should either run arq
* or keep anticipating a better request.
*/
static int as_can_anticipate(struct as_data *ad, struct as_rq *arq)
* Status is either ANTIC_OFF so start waiting,
* ANTIC_WAIT_REQ so continue waiting for request to finish
* or ANTIC_WAIT_NEXT so continue waiting for an acceptable request.
- *
*/
return 1;
}
-static void as_update_thinktime(struct as_data *ad, struct as_io_context *aic, unsigned long ttime)
-{
- /* fixed point: 1.0 == 1<<8 */
- if (aic->ttime_samples == 0) {
- ad->new_ttime_total = (7*ad->new_ttime_total + 256*ttime) / 8;
- ad->new_ttime_mean = ad->new_ttime_total / 256;
-
- ad->exit_prob = (7*ad->exit_prob)/8;
- }
- aic->ttime_samples = (7*aic->ttime_samples + 256) / 8;
- aic->ttime_total = (7*aic->ttime_total + 256*ttime) / 8;
- aic->ttime_mean = (aic->ttime_total + 128) / aic->ttime_samples;
-}
-
-static void as_update_seekdist(struct as_data *ad, struct as_io_context *aic, sector_t sdist)
-{
- u64 total;
-
- if (aic->seek_samples == 0) {
- ad->new_seek_total = (7*ad->new_seek_total + 256*(u64)sdist)/8;
- ad->new_seek_mean = ad->new_seek_total / 256;
- }
-
- /*
- * Don't allow the seek distance to get too large from the
- * odd fragment, pagein, etc
- */
- if (aic->seek_samples <= 60) /* second&third seek */
- sdist = min(sdist, (aic->seek_mean * 4) + 2*1024*1024);
- else
- sdist = min(sdist, (aic->seek_mean * 4) + 2*1024*64);
-
- aic->seek_samples = (7*aic->seek_samples + 256) / 8;
- aic->seek_total = (7*aic->seek_total + (u64)256*sdist) / 8;
- total = aic->seek_total + (aic->seek_samples/2);
- do_div(total, aic->seek_samples);
- aic->seek_mean = (sector_t)total;
-}
-
-/*
- * as_update_iohist keeps a decaying histogram of IO thinktimes, and
- * updates @aic->ttime_mean based on that. It is called when a new
- * request is queued.
- */
-static void as_update_iohist(struct as_data *ad, struct as_io_context *aic, struct request *rq)
-{
- struct as_rq *arq = RQ_DATA(rq);
- int data_dir = arq->is_sync;
- unsigned long thinktime;
- sector_t seek_dist;
-
- if (aic == NULL)
- return;
-
- if (data_dir == REQ_SYNC) {
- unsigned long in_flight = atomic_read(&aic->nr_queued)
- + atomic_read(&aic->nr_dispatched);
- spin_lock(&aic->lock);
- if (test_bit(AS_TASK_IORUNNING, &aic->state) ||
- test_bit(AS_TASK_IOSTARTED, &aic->state)) {
- /* Calculate read -> read thinktime */
- if (test_bit(AS_TASK_IORUNNING, &aic->state)
- && in_flight == 0) {
- thinktime = jiffies - aic->last_end_request;
- thinktime = min(thinktime, MAX_THINKTIME-1);
- } else
- thinktime = 0;
- 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;
- else
- seek_dist = aic->last_request_pos - rq->sector;
- as_update_seekdist(ad, aic, seek_dist);
- }
- aic->last_request_pos = rq->sector + rq->nr_sectors;
- set_bit(AS_TASK_IOSTARTED, &aic->state);
- spin_unlock(&aic->lock);
- }
-}
-
/*
* as_update_arq must be called whenever a request (arq) is added to
* the sort_list. This function keeps caches up to date, and checks if the
ad->next_arq[data_dir] = as_find_next_arq(ad, arq);
list_del_init(&arq->fifo);
- as_del_arq_hash(arq);
- as_del_arq_rb(ad, arq);
+ as_del_arq_rb(ad, rq);
}
/*
struct request *rq = arq->request;
const int data_dir = arq->is_sync;
- BUG_ON(!ON_RB(&arq->rb_node));
+ BUG_ON(RB_EMPTY_NODE(&rq->rb_node));
as_antic_stop(ad);
ad->antic_status = ANTIC_OFF;
/*
* take it off the sort and fifo list, add to dispatch queue
*/
- while (!list_empty(&rq->queuelist)) {
- struct request *__rq = list_entry_rq(rq->queuelist.next);
- struct as_rq *__arq = RQ_DATA(__rq);
-
- list_del(&__rq->queuelist);
-
- elv_dispatch_add_tail(ad->q, __rq);
-
- if (__arq->io_context && __arq->io_context->aic)
- atomic_inc(&__arq->io_context->aic->nr_dispatched);
-
- WARN_ON(__arq->state != AS_RQ_QUEUED);
- __arq->state = AS_RQ_DISPATCHED;
-
- ad->nr_dispatched++;
- }
-
as_remove_queued_request(ad->q, rq);
WARN_ON(arq->state != AS_RQ_QUEUED);
|| ad->changed_batch)
return 0;
- if (!(reads && writes && as_batch_expired(ad)) ) {
+ if (!(reads && writes && as_batch_expired(ad))) {
/*
* batch is still running or no reads or no writes
*/
*/
if (reads) {
- BUG_ON(RB_EMPTY(&ad->sort_list[REQ_SYNC]));
+ BUG_ON(RB_EMPTY_ROOT(&ad->sort_list[REQ_SYNC]));
if (writes && ad->batch_data_dir == REQ_SYNC)
/*
if (writes) {
dispatch_writes:
- BUG_ON(RB_EMPTY(&ad->sort_list[REQ_ASYNC]));
+ BUG_ON(RB_EMPTY_ROOT(&ad->sort_list[REQ_ASYNC]));
if (ad->batch_data_dir == REQ_SYNC) {
ad->changed_batch = 1;
}
/*
- * Add arq to a list behind alias
- */
-static inline void
-as_add_aliased_request(struct as_data *ad, struct as_rq *arq, struct as_rq *alias)
-{
- struct request *req = arq->request;
- struct list_head *insert = alias->request->queuelist.prev;
-
- /*
- * Transfer list of aliases
- */
- while (!list_empty(&req->queuelist)) {
- struct request *__rq = list_entry_rq(req->queuelist.next);
- struct as_rq *__arq = RQ_DATA(__rq);
-
- list_move_tail(&__rq->queuelist, &alias->request->queuelist);
-
- WARN_ON(__arq->state != AS_RQ_QUEUED);
- }
-
- /*
- * Another request with the same start sector on the rbtree.
- * Link this request to that sector. They are untangled in
- * as_move_to_dispatch
- */
- list_add(&arq->request->queuelist, insert);
-
- /*
- * Don't want to have to handle merges.
- */
- as_del_arq_hash(arq);
- arq->request->flags |= REQ_NOMERGE;
-}
-
-/*
* add arq to rbtree and fifo
*/
static void as_add_request(request_queue_t *q, struct request *rq)
{
struct as_data *ad = q->elevator->elevator_data;
struct as_rq *arq = RQ_DATA(rq);
- struct as_rq *alias;
int data_dir;
- if (arq->state != AS_RQ_PRESCHED) {
- printk("arq->state: %d\n", arq->state);
- WARN_ON(1);
- }
arq->state = AS_RQ_NEW;
if (rq_data_dir(arq->request) == READ
- || current->flags&PF_SYNCWRITE)
+ || (arq->request->cmd_flags & REQ_RW_SYNC))
arq->is_sync = 1;
else
arq->is_sync = 0;
atomic_inc(&arq->io_context->aic->nr_queued);
}
- alias = as_add_arq_rb(ad, arq);
- if (!alias) {
- /*
- * set expire time (only used for reads) and add to fifo list
- */
- arq->expires = jiffies + ad->fifo_expire[data_dir];
- list_add_tail(&arq->fifo, &ad->fifo_list[data_dir]);
-
- if (rq_mergeable(arq->request))
- as_add_arq_hash(ad, arq);
- as_update_arq(ad, arq); /* keep state machine up to date */
-
- } else {
- as_add_aliased_request(ad, arq, alias);
+ as_add_arq_rb(ad, rq);
- /*
- * have we been anticipating this request?
- * or does it come from the same process as the one we are
- * anticipating for?
- */
- if (ad->antic_status == ANTIC_WAIT_REQ
- || ad->antic_status == ANTIC_WAIT_NEXT) {
- if (as_can_break_anticipation(ad, arq))
- as_antic_stop(ad);
- }
- }
+ /*
+ * set expire time (only used for reads) and add to fifo list
+ */
+ arq->expires = jiffies + ad->fifo_expire[data_dir];
+ list_add_tail(&arq->fifo, &ad->fifo_list[data_dir]);
+ as_update_arq(ad, arq); /* keep state machine up to date */
arq->state = AS_RQ_QUEUED;
}
&& list_empty(&ad->fifo_list[REQ_SYNC]);
}
-static struct request *
-as_former_request(request_queue_t *q, struct request *rq)
-{
- struct as_rq *arq = RQ_DATA(rq);
- struct rb_node *rbprev = rb_prev(&arq->rb_node);
- struct request *ret = NULL;
-
- if (rbprev)
- ret = rb_entry_arq(rbprev)->request;
-
- return ret;
-}
-
-static struct request *
-as_latter_request(request_queue_t *q, struct request *rq)
-{
- struct as_rq *arq = RQ_DATA(rq);
- struct rb_node *rbnext = rb_next(&arq->rb_node);
- struct request *ret = NULL;
-
- if (rbnext)
- ret = rb_entry_arq(rbnext)->request;
-
- return ret;
-}
-
static int
as_merge(request_queue_t *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);
struct request *__rq;
- int ret;
-
- /*
- * see if the merge hash can satisfy a back merge
- */
- __rq = as_find_arq_hash(ad, bio->bi_sector);
- if (__rq) {
- BUG_ON(__rq->sector + __rq->nr_sectors != bio->bi_sector);
-
- if (elv_rq_merge_ok(__rq, bio)) {
- ret = ELEVATOR_BACK_MERGE;
- goto out;
- }
- }
/*
* check for front merge
*/
- __rq = as_find_arq_rb(ad, rb_key, bio_data_dir(bio));
- if (__rq) {
- BUG_ON(rb_key != rq_rb_key(__rq));
-
- if (elv_rq_merge_ok(__rq, bio)) {
- ret = ELEVATOR_FRONT_MERGE;
- goto out;
- }
+ __rq = elv_rb_find(&ad->sort_list[bio_data_dir(bio)], rb_key);
+ if (__rq && elv_rq_merge_ok(__rq, bio)) {
+ *req = __rq;
+ return ELEVATOR_FRONT_MERGE;
}
return ELEVATOR_NO_MERGE;
-out:
- if (ret) {
- if (rq_mergeable(__rq))
- as_hot_arq_hash(ad, RQ_DATA(__rq));
- }
- *req = __rq;
- return ret;
}
-static void as_merged_request(request_queue_t *q, struct request *req)
+static void as_merged_request(request_queue_t *q, struct request *req, int type)
{
struct as_data *ad = q->elevator->elevator_data;
- struct as_rq *arq = RQ_DATA(req);
-
- /*
- * hash always needs to be repositioned, key is end sector
- */
- as_del_arq_hash(arq);
- as_add_arq_hash(ad, arq);
/*
* if the merge was a front merge, we need to reposition request
*/
- if (rq_rb_key(req) != arq->rb_key) {
- struct as_rq *alias, *next_arq = NULL;
-
- if (ad->next_arq[arq->is_sync] == arq)
- next_arq = as_find_next_arq(ad, arq);
-
- /*
- * Note! We should really be moving any old aliased requests
- * off this request and try to insert them into the rbtree. We
- * currently don't bother. Ditto the next function.
- */
- as_del_arq_rb(ad, arq);
- if ((alias = as_add_arq_rb(ad, arq)) ) {
- list_del_init(&arq->fifo);
- as_add_aliased_request(ad, arq, alias);
- if (next_arq)
- ad->next_arq[arq->is_sync] = next_arq;
- }
+ if (type == ELEVATOR_FRONT_MERGE) {
+ as_del_arq_rb(ad, req);
+ as_add_arq_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,
- struct request *next)
+static void as_merged_requests(request_queue_t *q, struct request *req,
+ struct request *next)
{
- struct as_data *ad = q->elevator->elevator_data;
struct as_rq *arq = RQ_DATA(req);
struct as_rq *anext = RQ_DATA(next);
BUG_ON(!anext);
/*
- * reposition arq (this is the merged request) in hash, and in rbtree
- * in case of a front merge
- */
- as_del_arq_hash(arq);
- as_add_arq_hash(ad, arq);
-
- if (rq_rb_key(req) != arq->rb_key) {
- struct as_rq *alias, *next_arq = NULL;
-
- if (ad->next_arq[arq->is_sync] == arq)
- next_arq = as_find_next_arq(ad, arq);
-
- as_del_arq_rb(ad, arq);
- if ((alias = as_add_arq_rb(ad, arq)) ) {
- list_del_init(&arq->fifo);
- as_add_aliased_request(ad, arq, alias);
- if (next_arq)
- ad->next_arq[arq->is_sync] = next_arq;
- }
- }
-
- /*
* if anext expires before arq, assign its expire time to arq
* and move into anext position (anext will be deleted) in fifo
*/
}
/*
- * Transfer list of aliases
- */
- while (!list_empty(&next->queuelist)) {
- struct request *__rq = list_entry_rq(next->queuelist.next);
- struct as_rq *__arq = RQ_DATA(__rq);
-
- list_move_tail(&__rq->queuelist, &req->queuelist);
-
- WARN_ON(__arq->state != AS_RQ_QUEUED);
- }
-
- /*
* kill knowledge of next, this one is a goner
*/
as_remove_queued_request(q, next);
if (arq) {
memset(arq, 0, sizeof(*arq));
- RB_CLEAR(&arq->rb_node);
arq->request = rq;
arq->state = AS_RQ_PRESCHED;
arq->io_context = NULL;
- INIT_LIST_HEAD(&arq->hash);
- arq->on_hash = 0;
INIT_LIST_HEAD(&arq->fifo);
rq->elevator_private = arq;
return 0;
mempool_destroy(ad->arq_pool);
put_io_context(ad->io_context);
- kfree(ad->hash);
kfree(ad);
}
* initialize elevator private data (as_data), and alloc a arq for
* each request on the free lists
*/
-static int as_init_queue(request_queue_t *q, elevator_t *e)
+static void *as_init_queue(request_queue_t *q, elevator_t *e)
{
struct as_data *ad;
- int i;
if (!arq_pool)
- return -ENOMEM;
+ return NULL;
ad = kmalloc_node(sizeof(*ad), GFP_KERNEL, q->node);
if (!ad)
- return -ENOMEM;
+ return NULL;
memset(ad, 0, sizeof(*ad));
ad->q = q; /* Identify what queue the data belongs to */
- ad->hash = kmalloc_node(sizeof(struct list_head)*AS_HASH_ENTRIES,
- GFP_KERNEL, q->node);
- if (!ad->hash) {
- kfree(ad);
- return -ENOMEM;
- }
-
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->hash);
kfree(ad);
- return -ENOMEM;
+ return NULL;
}
/* anticipatory scheduling helpers */
init_timer(&ad->antic_timer);
INIT_WORK(&ad->antic_work, as_work_handler, q);
- for (i = 0; i < AS_HASH_ENTRIES; i++)
- INIT_LIST_HEAD(&ad->hash[i]);
-
INIT_LIST_HEAD(&ad->fifo_list[REQ_SYNC]);
INIT_LIST_HEAD(&ad->fifo_list[REQ_ASYNC]);
ad->sort_list[REQ_SYNC] = RB_ROOT;
ad->antic_expire = default_antic_expire;
ad->batch_expire[REQ_SYNC] = default_read_batch_expire;
ad->batch_expire[REQ_ASYNC] = default_write_batch_expire;
- e->elevator_data = ad;
ad->current_batch_expires = jiffies + ad->batch_expire[REQ_SYNC];
ad->write_batch_count = ad->batch_expire[REQ_ASYNC] / 10;
if (ad->write_batch_count < 2)
ad->write_batch_count = 2;
- return 0;
+ return ad;
}
/*
* sysfs parts below
*/
-struct as_fs_entry {
- struct attribute attr;
- ssize_t (*show)(struct as_data *, char *);
- ssize_t (*store)(struct as_data *, const char *, size_t);
-};
static ssize_t
as_var_show(unsigned int var, char *page)
return count;
}
-static ssize_t as_est_show(struct as_data *ad, char *page)
+static ssize_t est_time_show(elevator_t *e, char *page)
{
+ struct as_data *ad = e->elevator_data;
int pos = 0;
- pos += sprintf(page+pos, "%lu %% exit probability\n", 100*ad->exit_prob/256);
+ pos += sprintf(page+pos, "%lu %% exit probability\n",
+ 100*ad->exit_prob/256);
+ pos += sprintf(page+pos, "%lu %% probability of exiting without a "
+ "cooperating process submitting IO\n",
+ 100*ad->exit_no_coop/256);
pos += sprintf(page+pos, "%lu ms new thinktime\n", ad->new_ttime_mean);
- pos += sprintf(page+pos, "%llu sectors new seek distance\n", (unsigned long long)ad->new_seek_mean);
+ pos += sprintf(page+pos, "%llu sectors new seek distance\n",
+ (unsigned long long)ad->new_seek_mean);
return pos;
}
#define SHOW_FUNCTION(__FUNC, __VAR) \
-static ssize_t __FUNC(struct as_data *ad, char *page) \
+static ssize_t __FUNC(elevator_t *e, char *page) \
{ \
+ struct as_data *ad = e->elevator_data; \
return as_var_show(jiffies_to_msecs((__VAR)), (page)); \
}
-SHOW_FUNCTION(as_readexpire_show, ad->fifo_expire[REQ_SYNC]);
-SHOW_FUNCTION(as_writeexpire_show, ad->fifo_expire[REQ_ASYNC]);
-SHOW_FUNCTION(as_anticexpire_show, ad->antic_expire);
-SHOW_FUNCTION(as_read_batchexpire_show, ad->batch_expire[REQ_SYNC]);
-SHOW_FUNCTION(as_write_batchexpire_show, ad->batch_expire[REQ_ASYNC]);
+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_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]);
#undef SHOW_FUNCTION
#define STORE_FUNCTION(__FUNC, __PTR, MIN, MAX) \
-static ssize_t __FUNC(struct as_data *ad, const char *page, size_t count) \
+static ssize_t __FUNC(elevator_t *e, const char *page, size_t count) \
{ \
- int ret = as_var_store(__PTR, (page), count); \
+ struct as_data *ad = e->elevator_data; \
+ int ret = as_var_store(__PTR, (page), count); \
if (*(__PTR) < (MIN)) \
*(__PTR) = (MIN); \
else if (*(__PTR) > (MAX)) \
*(__PTR) = msecs_to_jiffies(*(__PTR)); \
return ret; \
}
-STORE_FUNCTION(as_readexpire_store, &ad->fifo_expire[REQ_SYNC], 0, INT_MAX);
-STORE_FUNCTION(as_writeexpire_store, &ad->fifo_expire[REQ_ASYNC], 0, INT_MAX);
-STORE_FUNCTION(as_anticexpire_store, &ad->antic_expire, 0, INT_MAX);
-STORE_FUNCTION(as_read_batchexpire_store,
+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_antic_expire_store, &ad->antic_expire, 0, INT_MAX);
+STORE_FUNCTION(as_read_batch_expire_store,
&ad->batch_expire[REQ_SYNC], 0, INT_MAX);
-STORE_FUNCTION(as_write_batchexpire_store,
+STORE_FUNCTION(as_write_batch_expire_store,
&ad->batch_expire[REQ_ASYNC], 0, INT_MAX);
#undef STORE_FUNCTION
-static struct as_fs_entry as_est_entry = {
- .attr = {.name = "est_time", .mode = S_IRUGO },
- .show = as_est_show,
-};
-static struct as_fs_entry as_readexpire_entry = {
- .attr = {.name = "read_expire", .mode = S_IRUGO | S_IWUSR },
- .show = as_readexpire_show,
- .store = as_readexpire_store,
-};
-static struct as_fs_entry as_writeexpire_entry = {
- .attr = {.name = "write_expire", .mode = S_IRUGO | S_IWUSR },
- .show = as_writeexpire_show,
- .store = as_writeexpire_store,
-};
-static struct as_fs_entry as_anticexpire_entry = {
- .attr = {.name = "antic_expire", .mode = S_IRUGO | S_IWUSR },
- .show = as_anticexpire_show,
- .store = as_anticexpire_store,
-};
-static struct as_fs_entry as_read_batchexpire_entry = {
- .attr = {.name = "read_batch_expire", .mode = S_IRUGO | S_IWUSR },
- .show = as_read_batchexpire_show,
- .store = as_read_batchexpire_store,
-};
-static struct as_fs_entry as_write_batchexpire_entry = {
- .attr = {.name = "write_batch_expire", .mode = S_IRUGO | S_IWUSR },
- .show = as_write_batchexpire_show,
- .store = as_write_batchexpire_store,
-};
-
-static struct attribute *default_attrs[] = {
- &as_est_entry.attr,
- &as_readexpire_entry.attr,
- &as_writeexpire_entry.attr,
- &as_anticexpire_entry.attr,
- &as_read_batchexpire_entry.attr,
- &as_write_batchexpire_entry.attr,
- NULL,
-};
-
-#define to_as(atr) container_of((atr), struct as_fs_entry, attr)
-
-static ssize_t
-as_attr_show(struct kobject *kobj, struct attribute *attr, char *page)
-{
- elevator_t *e = container_of(kobj, elevator_t, kobj);
- struct as_fs_entry *entry = to_as(attr);
-
- if (!entry->show)
- return -EIO;
-
- return entry->show(e->elevator_data, page);
-}
-
-static ssize_t
-as_attr_store(struct kobject *kobj, struct attribute *attr,
- const char *page, size_t length)
-{
- elevator_t *e = container_of(kobj, elevator_t, kobj);
- struct as_fs_entry *entry = to_as(attr);
-
- if (!entry->store)
- return -EIO;
-
- return entry->store(e->elevator_data, page, length);
-}
-
-static struct sysfs_ops as_sysfs_ops = {
- .show = as_attr_show,
- .store = as_attr_store,
-};
-
-static struct kobj_type as_ktype = {
- .sysfs_ops = &as_sysfs_ops,
- .default_attrs = default_attrs,
+#define AS_ATTR(name) \
+ __ATTR(name, S_IRUGO|S_IWUSR, as_##name##_show, as_##name##_store)
+
+static struct elv_fs_entry as_attrs[] = {
+ __ATTR_RO(est_time),
+ AS_ATTR(read_expire),
+ AS_ATTR(write_expire),
+ AS_ATTR(antic_expire),
+ AS_ATTR(read_batch_expire),
+ AS_ATTR(write_batch_expire),
+ __ATTR_NULL
};
static struct elevator_type iosched_as = {
.elevator_deactivate_req_fn = as_deactivate_request,
.elevator_queue_empty_fn = as_queue_empty,
.elevator_completed_req_fn = as_completed_request,
- .elevator_former_req_fn = as_former_request,
- .elevator_latter_req_fn = as_latter_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,
+ .trim = as_trim,
},
- .elevator_ktype = &as_ktype,
+ .elevator_attrs = as_attrs,
.elevator_name = "anticipatory",
.elevator_owner = THIS_MODULE,
};
static void __exit as_exit(void)
{
+ DECLARE_COMPLETION(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);
+ synchronize_rcu();
kmem_cache_destroy(arq_pool);
}