From 5db5d64277bf390056b1a87d0bb288c8b8553f96 Mon Sep 17 00:00:00 2001 From: Corrado Zoccolo Date: Mon, 26 Oct 2009 22:44:04 +0100 Subject: cfq-iosched: adapt slice to number of processes doing I/O When the number of processes performing I/O concurrently increases, a fixed time slice per process will cause large latencies. This patch, if low_latency mode is enabled, will scale the time slice assigned to each process according to a 300ms target latency. In order to keep fairness among processes: * The number of active processes is computed using a special form of running average, that quickly follows sudden increases (to keep latency low), and decrease slowly (to have fairness in spite of rapid decreases of this value). To safeguard sequential bandwidth, we impose a minimum time slice (computed using 2*cfq_slice_idle as base, adjusted according to priority and async-ness). Signed-off-by: Corrado Zoccolo Signed-off-by: Jens Axboe --- block/cfq-iosched.c | 53 +++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 51 insertions(+), 2 deletions(-) (limited to 'block/cfq-iosched.c') diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c index 418da9a49bb0..97d946585bc3 100644 --- a/block/cfq-iosched.c +++ b/block/cfq-iosched.c @@ -27,6 +27,8 @@ static const int cfq_slice_sync = HZ / 10; static int cfq_slice_async = HZ / 25; static const int cfq_slice_async_rq = 2; static int cfq_slice_idle = HZ / 125; +static const int cfq_target_latency = HZ * 3/10; /* 300 ms */ +static const int cfq_hist_divisor = 4; /* * offset from end of service tree @@ -148,6 +150,8 @@ struct cfq_data { struct rb_root prio_trees[CFQ_PRIO_LISTS]; unsigned int busy_queues; + unsigned int busy_rt_queues; + unsigned int busy_queues_avg[2]; int rq_in_driver[2]; int sync_flight; @@ -315,10 +319,52 @@ cfq_prio_to_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq) return cfq_prio_slice(cfqd, cfq_cfqq_sync(cfqq), cfqq->ioprio); } +/* + * get averaged number of queues of RT/BE priority. + * average is updated, with a formula that gives more weight to higher numbers, + * to quickly follows sudden increases and decrease slowly + */ + +static inline unsigned +cfq_get_avg_queues(struct cfq_data *cfqd, bool rt) { + unsigned min_q, max_q; + unsigned mult = cfq_hist_divisor - 1; + unsigned round = cfq_hist_divisor / 2; + unsigned busy = cfqd->busy_rt_queues; + + if (!rt) + busy = cfqd->busy_queues - cfqd->busy_rt_queues; + + min_q = min(cfqd->busy_queues_avg[rt], busy); + max_q = max(cfqd->busy_queues_avg[rt], busy); + cfqd->busy_queues_avg[rt] = (mult * max_q + min_q + round) / + cfq_hist_divisor; + return cfqd->busy_queues_avg[rt]; +} + static inline void cfq_set_prio_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq) { - cfqq->slice_end = cfq_prio_to_slice(cfqd, cfqq) + jiffies; + unsigned slice = cfq_prio_to_slice(cfqd, cfqq); + if (cfqd->cfq_latency) { + /* interested queues (we consider only the ones with the same + * priority class) */ + unsigned iq = cfq_get_avg_queues(cfqd, cfq_class_rt(cfqq)); + unsigned sync_slice = cfqd->cfq_slice[1]; + unsigned expect_latency = sync_slice * iq; + if (expect_latency > cfq_target_latency) { + unsigned base_low_slice = 2 * cfqd->cfq_slice_idle; + /* scale low_slice according to IO priority + * and sync vs async */ + unsigned low_slice = + min(slice, base_low_slice * slice / sync_slice); + /* the adapted slice value is scaled to fit all iqs + * into the target latency */ + slice = max(slice * cfq_target_latency / expect_latency, + low_slice); + } + } + cfqq->slice_end = jiffies + slice; cfq_log_cfqq(cfqd, cfqq, "set_slice=%lu", cfqq->slice_end - jiffies); } @@ -669,7 +715,8 @@ static void cfq_add_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq) BUG_ON(cfq_cfqq_on_rr(cfqq)); cfq_mark_cfqq_on_rr(cfqq); cfqd->busy_queues++; - + if (cfq_class_rt(cfqq)) + cfqd->busy_rt_queues++; cfq_resort_rr_list(cfqd, cfqq); } @@ -692,6 +739,8 @@ static void cfq_del_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq) BUG_ON(!cfqd->busy_queues); cfqd->busy_queues--; + if (cfq_class_rt(cfqq)) + cfqd->busy_rt_queues--; } /* -- cgit v1.2.3 From aa6f6a3de18131348f70951efb2c56d806033e09 Mon Sep 17 00:00:00 2001 From: Corrado Zoccolo Date: Mon, 26 Oct 2009 22:44:33 +0100 Subject: cfq-iosched: preparation to handle multiple service trees We embed a pointer to the service tree in each queue, to handle multiple service trees easily. Service trees are enriched with a counter. cfq_add_rq_rb is invoked after putting the rq in the fifo, to ensure that all fields in rq are properly initialized. Signed-off-by: Corrado Zoccolo Signed-off-by: Jens Axboe --- block/cfq-iosched.c | 30 +++++++++++++++++++----------- 1 file changed, 19 insertions(+), 11 deletions(-) (limited to 'block/cfq-iosched.c') diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c index 97d946585bc3..c95c69e199f4 100644 --- a/block/cfq-iosched.c +++ b/block/cfq-iosched.c @@ -75,8 +75,9 @@ static DEFINE_SPINLOCK(ioc_gone_lock); struct cfq_rb_root { struct rb_root rb; struct rb_node *left; + unsigned count; }; -#define CFQ_RB_ROOT (struct cfq_rb_root) { RB_ROOT, NULL, } +#define CFQ_RB_ROOT (struct cfq_rb_root) { RB_ROOT, NULL, 0, } /* * Per process-grouping structure @@ -128,6 +129,7 @@ struct cfq_queue { pid_t pid; + struct cfq_rb_root *service_tree; struct cfq_queue *new_cfqq; }; @@ -503,6 +505,7 @@ static void cfq_rb_erase(struct rb_node *n, struct cfq_rb_root *root) if (root->left == n) root->left = NULL; rb_erase_init(n, &root->rb); + --root->count; } /* @@ -553,11 +556,12 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq, struct rb_node **p, *parent; struct cfq_queue *__cfqq; unsigned long rb_key; + struct cfq_rb_root *service_tree = &cfqd->service_tree; int left; if (cfq_class_idle(cfqq)) { rb_key = CFQ_IDLE_DELAY; - parent = rb_last(&cfqd->service_tree.rb); + parent = rb_last(&service_tree->rb); if (parent && parent != &cfqq->rb_node) { __cfqq = rb_entry(parent, struct cfq_queue, rb_node); rb_key += __cfqq->rb_key; @@ -575,7 +579,7 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq, cfqq->slice_resid = 0; } else { rb_key = -HZ; - __cfqq = cfq_rb_first(&cfqd->service_tree); + __cfqq = cfq_rb_first(service_tree); rb_key += __cfqq ? __cfqq->rb_key : jiffies; } @@ -586,12 +590,14 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq, if (rb_key == cfqq->rb_key) return; - cfq_rb_erase(&cfqq->rb_node, &cfqd->service_tree); + cfq_rb_erase(&cfqq->rb_node, cfqq->service_tree); + cfqq->service_tree = NULL; } left = 1; parent = NULL; - p = &cfqd->service_tree.rb.rb_node; + cfqq->service_tree = service_tree; + p = &service_tree->rb.rb_node; while (*p) { struct rb_node **n; @@ -623,11 +629,12 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq, } if (left) - cfqd->service_tree.left = &cfqq->rb_node; + service_tree->left = &cfqq->rb_node; cfqq->rb_key = rb_key; rb_link_node(&cfqq->rb_node, parent, p); - rb_insert_color(&cfqq->rb_node, &cfqd->service_tree.rb); + rb_insert_color(&cfqq->rb_node, &service_tree->rb); + service_tree->count++; } static struct cfq_queue * @@ -730,8 +737,10 @@ static void cfq_del_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq) BUG_ON(!cfq_cfqq_on_rr(cfqq)); cfq_clear_cfqq_on_rr(cfqq); - if (!RB_EMPTY_NODE(&cfqq->rb_node)) - cfq_rb_erase(&cfqq->rb_node, &cfqd->service_tree); + if (!RB_EMPTY_NODE(&cfqq->rb_node)) { + cfq_rb_erase(&cfqq->rb_node, cfqq->service_tree); + cfqq->service_tree = NULL; + } if (cfqq->p_root) { rb_erase(&cfqq->p_node, cfqq->p_root); cfqq->p_root = NULL; @@ -2292,10 +2301,9 @@ static void cfq_insert_request(struct request_queue *q, struct request *rq) cfq_log_cfqq(cfqd, cfqq, "insert_request"); cfq_init_prio_data(cfqq, RQ_CIC(rq)->ioc); - cfq_add_rq_rb(rq); - rq_set_fifo_time(rq, jiffies + cfqd->cfq_fifo_expire[rq_is_sync(rq)]); list_add_tail(&rq->queuelist, &cfqq->fifo); + cfq_add_rq_rb(rq); cfq_rq_enqueued(cfqd, cfqq, rq); } -- cgit v1.2.3 From c0324a020e5b351f100569b128715985f1023af8 Mon Sep 17 00:00:00 2001 From: Corrado Zoccolo Date: Tue, 27 Oct 2009 19:16:03 +0100 Subject: cfq-iosched: reimplement priorities using different service trees We use different service trees for different priority classes. This allows a simplification in the service tree insertion code, that no longer has to consider priority while walking the tree. Signed-off-by: Corrado Zoccolo Signed-off-by: Jens Axboe --- block/cfq-iosched.c | 116 +++++++++++++++++++++++++++++++++++++--------------- 1 file changed, 82 insertions(+), 34 deletions(-) (limited to 'block/cfq-iosched.c') diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c index c95c69e199f4..6e5c3d715ebe 100644 --- a/block/cfq-iosched.c +++ b/block/cfq-iosched.c @@ -133,6 +133,16 @@ struct cfq_queue { struct cfq_queue *new_cfqq; }; +/* + * Index in the service_trees. + * IDLE is handled separately, so it has negative index + */ +enum wl_prio_t { + IDLE_WORKLOAD = -1, + BE_WORKLOAD = 0, + RT_WORKLOAD = 1 +}; + /* * Per block device queue structure */ @@ -140,9 +150,15 @@ struct cfq_data { struct request_queue *queue; /* - * rr list of queues with requests and the count of them + * rr lists of queues with requests, onle rr for each priority class. + * Counts are embedded in the cfq_rb_root + */ + struct cfq_rb_root service_trees[2]; + struct cfq_rb_root service_tree_idle; + /* + * The priority currently being served */ - struct cfq_rb_root service_tree; + enum wl_prio_t serving_prio; /* * Each priority tree is sorted by next_request position. These @@ -152,7 +168,6 @@ struct cfq_data { struct rb_root prio_trees[CFQ_PRIO_LISTS]; unsigned int busy_queues; - unsigned int busy_rt_queues; unsigned int busy_queues_avg[2]; int rq_in_driver[2]; @@ -205,6 +220,15 @@ struct cfq_data { unsigned long last_end_sync_rq; }; +static struct cfq_rb_root *service_tree_for(enum wl_prio_t prio, + struct cfq_data *cfqd) +{ + if (prio == IDLE_WORKLOAD) + return &cfqd->service_tree_idle; + + return &cfqd->service_trees[prio]; +} + enum cfqq_state_flags { CFQ_CFQQ_FLAG_on_rr = 0, /* on round-robin busy list */ CFQ_CFQQ_FLAG_wait_request, /* waiting for a request */ @@ -249,6 +273,23 @@ CFQ_CFQQ_FNS(coop); #define cfq_log(cfqd, fmt, args...) \ blk_add_trace_msg((cfqd)->queue, "cfq " fmt, ##args) +static inline enum wl_prio_t cfqq_prio(struct cfq_queue *cfqq) +{ + if (cfq_class_idle(cfqq)) + return IDLE_WORKLOAD; + if (cfq_class_rt(cfqq)) + return RT_WORKLOAD; + return BE_WORKLOAD; +} + +static inline int cfq_busy_queues_wl(enum wl_prio_t wl, struct cfq_data *cfqd) +{ + if (wl == IDLE_WORKLOAD) + return cfqd->service_tree_idle.count; + + return cfqd->service_trees[wl].count; +} + static void cfq_dispatch_insert(struct request_queue *, struct request *); static struct cfq_queue *cfq_get_queue(struct cfq_data *, bool, struct io_context *, gfp_t); @@ -332,10 +373,7 @@ cfq_get_avg_queues(struct cfq_data *cfqd, bool rt) { unsigned min_q, max_q; unsigned mult = cfq_hist_divisor - 1; unsigned round = cfq_hist_divisor / 2; - unsigned busy = cfqd->busy_rt_queues; - - if (!rt) - busy = cfqd->busy_queues - cfqd->busy_rt_queues; + unsigned busy = cfq_busy_queues_wl(rt, cfqd); min_q = min(cfqd->busy_queues_avg[rt], busy); max_q = max(cfqd->busy_queues_avg[rt], busy); @@ -546,7 +584,7 @@ static unsigned long cfq_slice_offset(struct cfq_data *cfqd, } /* - * The cfqd->service_tree holds all pending cfq_queue's that have + * The cfqd->service_trees holds all pending cfq_queue's that have * requests waiting to be processed. It is sorted in the order that * we will service the queues. */ @@ -556,9 +594,10 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq, struct rb_node **p, *parent; struct cfq_queue *__cfqq; unsigned long rb_key; - struct cfq_rb_root *service_tree = &cfqd->service_tree; + struct cfq_rb_root *service_tree; int left; + service_tree = service_tree_for(cfqq_prio(cfqq), cfqd); if (cfq_class_idle(cfqq)) { rb_key = CFQ_IDLE_DELAY; parent = rb_last(&service_tree->rb); @@ -587,7 +626,8 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq, /* * same position, nothing more to do */ - if (rb_key == cfqq->rb_key) + if (rb_key == cfqq->rb_key && + cfqq->service_tree == service_tree) return; cfq_rb_erase(&cfqq->rb_node, cfqq->service_tree); @@ -605,25 +645,14 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq, __cfqq = rb_entry(parent, struct cfq_queue, rb_node); /* - * sort RT queues first, we always want to give - * preference to them. IDLE queues goes to the back. - * after that, sort on the next service time. + * sort by key, that represents service time. */ - if (cfq_class_rt(cfqq) > cfq_class_rt(__cfqq)) + if (time_before(rb_key, __cfqq->rb_key)) n = &(*p)->rb_left; - else if (cfq_class_rt(cfqq) < cfq_class_rt(__cfqq)) - n = &(*p)->rb_right; - else if (cfq_class_idle(cfqq) < cfq_class_idle(__cfqq)) - n = &(*p)->rb_left; - else if (cfq_class_idle(cfqq) > cfq_class_idle(__cfqq)) - n = &(*p)->rb_right; - else if (time_before(rb_key, __cfqq->rb_key)) - n = &(*p)->rb_left; - else + else { n = &(*p)->rb_right; - - if (n == &(*p)->rb_right) left = 0; + } p = n; } @@ -722,8 +751,7 @@ static void cfq_add_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq) BUG_ON(cfq_cfqq_on_rr(cfqq)); cfq_mark_cfqq_on_rr(cfqq); cfqd->busy_queues++; - if (cfq_class_rt(cfqq)) - cfqd->busy_rt_queues++; + cfq_resort_rr_list(cfqd, cfqq); } @@ -748,8 +776,6 @@ static void cfq_del_cfqq_rr(struct cfq_data *cfqd, struct cfq_queue *cfqq) BUG_ON(!cfqd->busy_queues); cfqd->busy_queues--; - if (cfq_class_rt(cfqq)) - cfqd->busy_rt_queues--; } /* @@ -1003,10 +1029,12 @@ static inline void cfq_slice_expired(struct cfq_data *cfqd, bool timed_out) */ static struct cfq_queue *cfq_get_next_queue(struct cfq_data *cfqd) { - if (RB_EMPTY_ROOT(&cfqd->service_tree.rb)) - return NULL; + struct cfq_rb_root *service_tree = + service_tree_for(cfqd->serving_prio, cfqd); - return cfq_rb_first(&cfqd->service_tree); + if (RB_EMPTY_ROOT(&service_tree->rb)) + return NULL; + return cfq_rb_first(service_tree); } /* @@ -1123,6 +1151,12 @@ static struct cfq_queue *cfq_close_cooperator(struct cfq_data *cfqd, if (CFQQ_SEEKY(cfqq)) return NULL; + /* + * Do not merge queues of different priority classes + */ + if (cfq_class_rt(cfqq) != cfq_class_rt(cur_cfqq)) + return NULL; + return cfqq; } @@ -1336,6 +1370,14 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd) expire: cfq_slice_expired(cfqd, 0); new_queue: + if (!new_cfqq) { + if (cfq_busy_queues_wl(RT_WORKLOAD, cfqd)) + cfqd->serving_prio = RT_WORKLOAD; + else if (cfq_busy_queues_wl(BE_WORKLOAD, cfqd)) + cfqd->serving_prio = BE_WORKLOAD; + else + cfqd->serving_prio = IDLE_WORKLOAD; + } cfqq = cfq_set_active_queue(cfqd, new_cfqq); keep_queue: return cfqq; @@ -1362,8 +1404,12 @@ static int cfq_forced_dispatch(struct cfq_data *cfqd) { struct cfq_queue *cfqq; int dispatched = 0; + int i; + for (i = 0; i < 2; ++i) + while ((cfqq = cfq_rb_first(&cfqd->service_trees[i])) != NULL) + dispatched += __cfq_forced_dispatch_cfqq(cfqq); - while ((cfqq = cfq_rb_first(&cfqd->service_tree)) != NULL) + while ((cfqq = cfq_rb_first(&cfqd->service_tree_idle)) != NULL) dispatched += __cfq_forced_dispatch_cfqq(cfqq); cfq_slice_expired(cfqd, 0); @@ -2710,7 +2756,9 @@ static void *cfq_init_queue(struct request_queue *q) if (!cfqd) return NULL; - cfqd->service_tree = CFQ_RB_ROOT; + for (i = 0; i < 2; ++i) + cfqd->service_trees[i] = CFQ_RB_ROOT; + cfqd->service_tree_idle = CFQ_RB_ROOT; /* * Not strictly needed (since RB_ROOT just clears the node and we -- cgit v1.2.3 From a6d44e982d3734583b3b4e1d36921af8cfd61fc0 Mon Sep 17 00:00:00 2001 From: Corrado Zoccolo Date: Mon, 26 Oct 2009 22:45:11 +0100 Subject: cfq-iosched: enable idling for last queue on priority class cfq can disable idling for queues in various circumstances. When workloads of different priorities are competing, if the higher priority queue has idling disabled, lower priority queues may steal its disk share. For example, in a scenario with an RT process performing seeky reads vs a BE process performing sequential reads, on an NCQ enabled hardware, with low_latency unset, the RT process will dispatch only the few pending requests every full slice of service for the BE process. The patch solves this issue by always performing idle on the last queue at a given priority class > idle. If the same process, or one that can pre-empt it (so at the same priority or higher), submits a new request within the idle window, the lower priority queue won't dispatch, saving the disk bandwidth for higher priority ones. Note: this doesn't touch the non_rotational + NCQ case (no hardware to test if this is a benefit in that case). Signed-off-by: Corrado Zoccolo Signed-off-by: Jens Axboe --- block/cfq-iosched.c | 34 +++++++++++++++++++++++++++++++--- 1 file changed, 31 insertions(+), 3 deletions(-) (limited to 'block/cfq-iosched.c') diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c index 6e5c3d715ebe..76afa3696894 100644 --- a/block/cfq-iosched.c +++ b/block/cfq-iosched.c @@ -1160,6 +1160,34 @@ static struct cfq_queue *cfq_close_cooperator(struct cfq_data *cfqd, return cfqq; } +/* + * Determine whether we should enforce idle window for this queue. + */ + +static bool cfq_should_idle(struct cfq_data *cfqd, struct cfq_queue *cfqq) +{ + enum wl_prio_t prio = cfqq_prio(cfqq); + struct cfq_rb_root *service_tree; + + /* We never do for idle class queues. */ + if (prio == IDLE_WORKLOAD) + return false; + + /* We do for queues that were marked with idle window flag. */ + if (cfq_cfqq_idle_window(cfqq)) + return true; + + /* + * Otherwise, we do only if they are the last ones + * in their service tree. + */ + service_tree = service_tree_for(prio, cfqd); + if (service_tree->count == 0) + return true; + + return (service_tree->count == 1 && cfq_rb_first(service_tree) == cfqq); +} + static void cfq_arm_slice_timer(struct cfq_data *cfqd) { struct cfq_queue *cfqq = cfqd->active_queue; @@ -1180,7 +1208,7 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd) /* * idle is disabled, either manually or by past process history */ - if (!cfqd->cfq_slice_idle || !cfq_cfqq_idle_window(cfqq)) + if (!cfqd->cfq_slice_idle || !cfq_should_idle(cfqd, cfqq)) return; /* @@ -1362,7 +1390,7 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd) * conditions to happen (or time out) before selecting a new queue. */ if (timer_pending(&cfqd->idle_slice_timer) || - (cfqq->dispatched && cfq_cfqq_idle_window(cfqq))) { + (cfqq->dispatched && cfq_should_idle(cfqd, cfqq))) { cfqq = NULL; goto keep_queue; } @@ -1427,7 +1455,7 @@ static bool cfq_may_dispatch(struct cfq_data *cfqd, struct cfq_queue *cfqq) /* * Drain async requests before we start sync IO */ - if (cfq_cfqq_idle_window(cfqq) && cfqd->rq_in_driver[BLK_RW_ASYNC]) + if (cfq_should_idle(cfqd, cfqq) && cfqd->rq_in_driver[BLK_RW_ASYNC]) return false; /* -- cgit v1.2.3 From 718eee0579b802aabe3bafacf09d0a9b0830f1dd Mon Sep 17 00:00:00 2001 From: Corrado Zoccolo Date: Mon, 26 Oct 2009 22:45:29 +0100 Subject: cfq-iosched: fairness for sync no-idle queues Currently no-idle queues in cfq are not serviced fairly: even if they can only dispatch a small number of requests at a time, they have to compete with idling queues to be serviced, experiencing large latencies. We should notice, instead, that no-idle queues are the ones that would benefit most from having low latency, in fact they are any of: * processes with large think times (e.g. interactive ones like file managers) * seeky (e.g. programs faulting in their code at startup) * or marked as no-idle from upper levels, to improve latencies of those requests. This patch improves the fairness and latency for those queues, by: * separating sync idle, sync no-idle and async queues in separate service_trees, for each priority * service all no-idle queues together * and idling when the last no-idle queue has been serviced, to anticipate for more no-idle work * the timeslices allotted for idle and no-idle service_trees are computed proportionally to the number of processes in each set. Servicing all no-idle queues together should have a performance boost for NCQ-capable drives, without compromising fairness. Signed-off-by: Corrado Zoccolo Signed-off-by: Jens Axboe --- block/cfq-iosched.c | 200 +++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 168 insertions(+), 32 deletions(-) (limited to 'block/cfq-iosched.c') diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c index 76afa3696894..859f534ae9ef 100644 --- a/block/cfq-iosched.c +++ b/block/cfq-iosched.c @@ -134,7 +134,7 @@ struct cfq_queue { }; /* - * Index in the service_trees. + * First index in the service_trees. * IDLE is handled separately, so it has negative index */ enum wl_prio_t { @@ -143,6 +143,16 @@ enum wl_prio_t { RT_WORKLOAD = 1 }; +/* + * Second index in the service_trees. + */ +enum wl_type_t { + ASYNC_WORKLOAD = 0, + SYNC_NOIDLE_WORKLOAD = 1, + SYNC_WORKLOAD = 2 +}; + + /* * Per block device queue structure */ @@ -153,12 +163,14 @@ struct cfq_data { * rr lists of queues with requests, onle rr for each priority class. * Counts are embedded in the cfq_rb_root */ - struct cfq_rb_root service_trees[2]; + struct cfq_rb_root service_trees[2][3]; struct cfq_rb_root service_tree_idle; /* * The priority currently being served */ enum wl_prio_t serving_prio; + enum wl_type_t serving_type; + unsigned long workload_expires; /* * Each priority tree is sorted by next_request position. These @@ -221,12 +233,13 @@ struct cfq_data { }; static struct cfq_rb_root *service_tree_for(enum wl_prio_t prio, + enum wl_type_t type, struct cfq_data *cfqd) { if (prio == IDLE_WORKLOAD) return &cfqd->service_tree_idle; - return &cfqd->service_trees[prio]; + return &cfqd->service_trees[prio][type]; } enum cfqq_state_flags { @@ -282,12 +295,24 @@ static inline enum wl_prio_t cfqq_prio(struct cfq_queue *cfqq) return BE_WORKLOAD; } + +static enum wl_type_t cfqq_type(struct cfq_queue *cfqq) +{ + if (!cfq_cfqq_sync(cfqq)) + return ASYNC_WORKLOAD; + if (!cfq_cfqq_idle_window(cfqq)) + return SYNC_NOIDLE_WORKLOAD; + return SYNC_WORKLOAD; +} + static inline int cfq_busy_queues_wl(enum wl_prio_t wl, struct cfq_data *cfqd) { if (wl == IDLE_WORKLOAD) return cfqd->service_tree_idle.count; - return cfqd->service_trees[wl].count; + return cfqd->service_trees[wl][ASYNC_WORKLOAD].count + + cfqd->service_trees[wl][SYNC_NOIDLE_WORKLOAD].count + + cfqd->service_trees[wl][SYNC_WORKLOAD].count; } static void cfq_dispatch_insert(struct request_queue *, struct request *); @@ -597,7 +622,7 @@ static void cfq_service_tree_add(struct cfq_data *cfqd, struct cfq_queue *cfqq, struct cfq_rb_root *service_tree; int left; - service_tree = service_tree_for(cfqq_prio(cfqq), cfqd); + service_tree = service_tree_for(cfqq_prio(cfqq), cfqq_type(cfqq), cfqd); if (cfq_class_idle(cfqq)) { rb_key = CFQ_IDLE_DELAY; parent = rb_last(&service_tree->rb); @@ -1030,7 +1055,7 @@ static inline void cfq_slice_expired(struct cfq_data *cfqd, bool timed_out) static struct cfq_queue *cfq_get_next_queue(struct cfq_data *cfqd) { struct cfq_rb_root *service_tree = - service_tree_for(cfqd->serving_prio, cfqd); + service_tree_for(cfqd->serving_prio, cfqd->serving_type, cfqd); if (RB_EMPTY_ROOT(&service_tree->rb)) return NULL; @@ -1167,7 +1192,7 @@ static struct cfq_queue *cfq_close_cooperator(struct cfq_data *cfqd, static bool cfq_should_idle(struct cfq_data *cfqd, struct cfq_queue *cfqq) { enum wl_prio_t prio = cfqq_prio(cfqq); - struct cfq_rb_root *service_tree; + struct cfq_rb_root *service_tree = cfqq->service_tree; /* We never do for idle class queues. */ if (prio == IDLE_WORKLOAD) @@ -1181,7 +1206,9 @@ static bool cfq_should_idle(struct cfq_data *cfqd, struct cfq_queue *cfqq) * Otherwise, we do only if they are the last ones * in their service tree. */ - service_tree = service_tree_for(prio, cfqd); + if (!service_tree) + service_tree = service_tree_for(prio, cfqq_type(cfqq), cfqd); + if (service_tree->count == 0) return true; @@ -1235,14 +1262,20 @@ static void cfq_arm_slice_timer(struct cfq_data *cfqd) cfq_mark_cfqq_wait_request(cfqq); - /* - * we don't want to idle for seeks, but we do want to allow - * fair distribution of slice time for a process doing back-to-back - * seeks. so allow a little bit of time for him to submit a new rq - */ sl = cfqd->cfq_slice_idle; - if (sample_valid(cfqq->seek_samples) && CFQQ_SEEKY(cfqq)) + /* are we servicing noidle tree, and there are more queues? + * non-rotational or NCQ: no idle + * non-NCQ rotational : very small idle, to allow + * fair distribution of slice time for a process doing back-to-back + * seeks. + */ + if (cfqd->serving_type == SYNC_NOIDLE_WORKLOAD && + service_tree_for(cfqd->serving_prio, SYNC_NOIDLE_WORKLOAD, cfqd) + ->count > 0) { + if (blk_queue_nonrot(cfqd->queue) || cfqd->hw_tag) + return; sl = min(sl, msecs_to_jiffies(CFQ_MIN_TT)); + } mod_timer(&cfqd->idle_slice_timer, jiffies + sl); cfq_log_cfqq(cfqd, cfqq, "arm_idle: %lu", sl); @@ -1346,6 +1379,106 @@ static void cfq_setup_merge(struct cfq_queue *cfqq, struct cfq_queue *new_cfqq) } } +static enum wl_type_t cfq_choose_wl(struct cfq_data *cfqd, enum wl_prio_t prio, + bool prio_changed) +{ + struct cfq_queue *queue; + int i; + bool key_valid = false; + unsigned long lowest_key = 0; + enum wl_type_t cur_best = SYNC_NOIDLE_WORKLOAD; + + if (prio_changed) { + /* + * When priorities switched, we prefer starting + * from SYNC_NOIDLE (first choice), or just SYNC + * over ASYNC + */ + if (service_tree_for(prio, cur_best, cfqd)->count) + return cur_best; + cur_best = SYNC_WORKLOAD; + if (service_tree_for(prio, cur_best, cfqd)->count) + return cur_best; + + return ASYNC_WORKLOAD; + } + + for (i = 0; i < 3; ++i) { + /* otherwise, select the one with lowest rb_key */ + queue = cfq_rb_first(service_tree_for(prio, i, cfqd)); + if (queue && + (!key_valid || time_before(queue->rb_key, lowest_key))) { + lowest_key = queue->rb_key; + cur_best = i; + key_valid = true; + } + } + + return cur_best; +} + +static void choose_service_tree(struct cfq_data *cfqd) +{ + enum wl_prio_t previous_prio = cfqd->serving_prio; + bool prio_changed; + unsigned slice; + unsigned count; + + /* Choose next priority. RT > BE > IDLE */ + if (cfq_busy_queues_wl(RT_WORKLOAD, cfqd)) + cfqd->serving_prio = RT_WORKLOAD; + else if (cfq_busy_queues_wl(BE_WORKLOAD, cfqd)) + cfqd->serving_prio = BE_WORKLOAD; + else { + cfqd->serving_prio = IDLE_WORKLOAD; + cfqd->workload_expires = jiffies + 1; + return; + } + + /* + * For RT and BE, we have to choose also the type + * (SYNC, SYNC_NOIDLE, ASYNC), and to compute a workload + * expiration time + */ + prio_changed = (cfqd->serving_prio != previous_prio); + count = service_tree_for(cfqd->serving_prio, cfqd->serving_type, cfqd) + ->count; + + /* + * If priority didn't change, check workload expiration, + * and that we still have other queues ready + */ + if (!prio_changed && count && + !time_after(jiffies, cfqd->workload_expires)) + return; + + /* otherwise select new workload type */ + cfqd->serving_type = + cfq_choose_wl(cfqd, cfqd->serving_prio, prio_changed); + count = service_tree_for(cfqd->serving_prio, cfqd->serving_type, cfqd) + ->count; + + /* + * the workload slice is computed as a fraction of target latency + * proportional to the number of queues in that workload, over + * all the queues in the same priority class + */ + slice = cfq_target_latency * count / + max_t(unsigned, cfqd->busy_queues_avg[cfqd->serving_prio], + cfq_busy_queues_wl(cfqd->serving_prio, cfqd)); + + if (cfqd->serving_type == ASYNC_WORKLOAD) + /* async workload slice is scaled down according to + * the sync/async slice ratio. */ + slice = slice * cfqd->cfq_slice[0] / cfqd->cfq_slice[1]; + else + /* sync workload slice is at least 2 * cfq_slice_idle */ + slice = max(slice, 2 * cfqd->cfq_slice_idle); + + slice = max_t(unsigned, slice, CFQ_MIN_TT); + cfqd->workload_expires = jiffies + slice; +} + /* * Select a queue for service. If we have a current active queue, * check whether to continue servicing it, or retrieve and set a new one. @@ -1398,14 +1531,13 @@ static struct cfq_queue *cfq_select_queue(struct cfq_data *cfqd) expire: cfq_slice_expired(cfqd, 0); new_queue: - if (!new_cfqq) { - if (cfq_busy_queues_wl(RT_WORKLOAD, cfqd)) - cfqd->serving_prio = RT_WORKLOAD; - else if (cfq_busy_queues_wl(BE_WORKLOAD, cfqd)) - cfqd->serving_prio = BE_WORKLOAD; - else - cfqd->serving_prio = IDLE_WORKLOAD; - } + /* + * Current queue expired. Check if we have to switch to a new + * service tree + */ + if (!new_cfqq) + choose_service_tree(cfqd); + cfqq = cfq_set_active_queue(cfqd, new_cfqq); keep_queue: return cfqq; @@ -1432,10 +1564,12 @@ static int cfq_forced_dispatch(struct cfq_data *cfqd) { struct cfq_queue *cfqq; int dispatched = 0; - int i; + int i, j; for (i = 0; i < 2; ++i) - while ((cfqq = cfq_rb_first(&cfqd->service_trees[i])) != NULL) - dispatched += __cfq_forced_dispatch_cfqq(cfqq); + for (j = 0; j < 3; ++j) + while ((cfqq = cfq_rb_first(&cfqd->service_trees[i][j])) + != NULL) + dispatched += __cfq_forced_dispatch_cfqq(cfqq); while ((cfqq = cfq_rb_first(&cfqd->service_tree_idle)) != NULL) dispatched += __cfq_forced_dispatch_cfqq(cfqq); @@ -2218,13 +2352,10 @@ cfq_update_idle_window(struct cfq_data *cfqd, struct cfq_queue *cfqq, enable_idle = old_idle = cfq_cfqq_idle_window(cfqq); if (!atomic_read(&cic->ioc->nr_tasks) || !cfqd->cfq_slice_idle || - (!cfqd->cfq_latency && cfqd->hw_tag && CFQQ_SEEKY(cfqq))) + (sample_valid(cfqq->seek_samples) && CFQQ_SEEKY(cfqq))) enable_idle = 0; else if (sample_valid(cic->ttime_samples)) { - unsigned int slice_idle = cfqd->cfq_slice_idle; - if (sample_valid(cfqq->seek_samples) && CFQQ_SEEKY(cfqq)) - slice_idle = msecs_to_jiffies(CFQ_MIN_TT); - if (cic->ttime_mean > slice_idle) + if (cic->ttime_mean > cfqd->cfq_slice_idle) enable_idle = 0; else enable_idle = 1; @@ -2262,6 +2393,10 @@ cfq_should_preempt(struct cfq_data *cfqd, struct cfq_queue *new_cfqq, if (cfq_class_idle(cfqq)) return true; + if (cfqd->serving_type == SYNC_NOIDLE_WORKLOAD + && new_cfqq->service_tree == cfqq->service_tree) + return true; + /* * if the new request is sync, but the currently running queue is * not, let the sync request have priority. @@ -2778,14 +2913,15 @@ static void cfq_exit_queue(struct elevator_queue *e) static void *cfq_init_queue(struct request_queue *q) { struct cfq_data *cfqd; - int i; + int i, j; cfqd = kmalloc_node(sizeof(*cfqd), GFP_KERNEL | __GFP_ZERO, q->node); if (!cfqd) return NULL; for (i = 0; i < 2; ++i) - cfqd->service_trees[i] = CFQ_RB_ROOT; + for (j = 0; j < 3; ++j) + cfqd->service_trees[i][j] = CFQ_RB_ROOT; cfqd->service_tree_idle = CFQ_RB_ROOT; /* -- cgit v1.2.3 From 5869619cb5b26754574375472fe54a390edf34c7 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Wed, 28 Oct 2009 09:27:07 +0100 Subject: cfq-iosched: fix style issue in cfq_get_avg_queues() Line breaks and bad brace placement. Signed-off-by: Jens Axboe --- block/cfq-iosched.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'block/cfq-iosched.c') diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c index 859f534ae9ef..aa00d8f2d0b0 100644 --- a/block/cfq-iosched.c +++ b/block/cfq-iosched.c @@ -393,8 +393,8 @@ cfq_prio_to_slice(struct cfq_data *cfqd, struct cfq_queue *cfqq) * to quickly follows sudden increases and decrease slowly */ -static inline unsigned -cfq_get_avg_queues(struct cfq_data *cfqd, bool rt) { +static inline unsigned cfq_get_avg_queues(struct cfq_data *cfqd, bool rt) +{ unsigned min_q, max_q; unsigned mult = cfq_hist_divisor - 1; unsigned round = cfq_hist_divisor / 2; -- cgit v1.2.3