ash/huawei-mrd-kernel
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
huawei-mrd-kernel/block/blk-throttle.c

3614 lines
97 KiB
C
Executable file
Raw Permalink Blame History

/*
* Interface for controlling IO bandwidth on a request queue
*
* Copyright (C) 2010 Vivek Goyal <vgoyal@redhat.com>
*/
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/blkdev.h>
#include <linux/bio.h>
#include <linux/blktrace_api.h>
#include <linux/blk-cgroup.h>
#include "blk.h"
#define SHARE_SHIFT (14)
#define MAX_SHARE (1 << SHARE_SHIFT)
#define MAX_INFLIGHT (128)
#define MIN_INFLIGHT (0)
#define QUANTUM_INQUEUE_MIN (1)
#define QUANTUM_INQUEUE_MAX (128)
#define QUANTUM_INQUEUE_DEF (16)
#define BW_SLICE_MIN (8ULL << 20) /* 8MB/s */
#define BW_SLICE_MAX (256ULL << 20) /* 256MB/s */
#define BW_SLICE_DEF (32ULL << 20) /* 32MB/s */
#define IOPS_SLICE_MIN (32)
#define IOPS_SLICE_MAX (4096)
#define IOPS_SLICE_DEF (256)
#define MAX_THROTL_DELAY_TM_MS (3000) /* 3s */
#ifdef CONFIG_HUAWEI_IO_TRACING
#include <iotrace/iotrace.h>
DEFINE_TRACE(block_throttle_weight)
DEFINE_TRACE(block_throttle_dispatch)
DEFINE_TRACE(block_throttle_iocost)
DEFINE_TRACE(block_throttle_limit_start)
DEFINE_TRACE(block_throttle_limit_end)
DEFINE_TRACE(block_throttle_bio_in)
DEFINE_TRACE(block_throttle_bio_out)
#endif
enum blk_throtl_weight_onoff_mode {
BLK_THROTL_WEIGHT_OFF,
BLK_THROTL_WEIGHT_ON_FS,
BLK_THROTL_WEIGHT_ON_BLK,
};
/* Max dispatch from a group in 1 round */
static int throtl_grp_quantum = 8;
/* Total max dispatch from all groups in one round */
static int throtl_quantum = 32;
/* Throttling is performed over 100ms slice and after that slice is renewed */
static unsigned long throtl_slice = HZ/10; /* 100 ms */
static struct blkcg_policy blkcg_policy_throtl;
/* A workqueue to queue throttle related work */
static struct workqueue_struct *kthrotld_workqueue;
/*lint -save -e64 -e570 -e651 -e708 -e785*/
static DEFINE_MUTEX(throtl_mode_lock);
/*lint -restore*/
int blk_throtl_weight_offon = BLK_THROTL_WEIGHT_OFF;
/*
* To implement hierarchical throttling, throtl_grps form a tree and bios
* are dispatched upwards level by level until they reach the top and get
* issued. When dispatching bios from the children and local group at each
* level, if the bios are dispatched into a single bio_list, there's a risk
* of a local or child group which can queue many bios at once filling up
* the list starving others.
*
* To avoid such starvation, dispatched bios are queued separately
* according to where they came from. When they are again dispatched to
* the parent, they're popped in round-robin order so that no single source
* hogs the dispatch window.
*
* throtl_qnode is used to keep the queued bios separated by their sources.
* Bios are queued to throtl_qnode which in turn is queued to
* throtl_service_queue and then dispatched in round-robin order.
*
* It's also used to track the reference counts on blkg's. A qnode always
* belongs to a throtl_grp and gets queued on itself or the parent, so
* incrementing the reference of the associated throtl_grp when a qnode is
* queued and decrementing when dequeued is enough to keep the whole blkg
* tree pinned while bios are in flight.
*/
struct throtl_qnode {
struct list_head node; /* service_queue->queued[] */
struct bio_list bios; /* queued bios */
struct throtl_grp *tg; /* tg this qnode belongs to */
};
struct throtl_service_queue {
struct throtl_service_queue *parent_sq; /* the parent service_queue */
/*
* Bios queued directly to this service_queue or dispatched from
* children throtl_grp's.
*/
struct list_head queued[2]; /* throtl_qnode [READ/WRITE] */
unsigned int nr_queued[2]; /* number of queued bios */
/*
* RB tree of active children throtl_grp's, which are sorted by
* their ->disptime.
*/
struct rb_root pending_tree; /* RB tree of active tgs */
struct rb_node *first_pending; /* first node in the tree */
unsigned int nr_pending; /* # queued in the tree */
unsigned long first_pending_disptime; /* disptime of the first tg */
struct timer_list pending_timer; /* fires on first_pending_disptime */
unsigned int children_weight; /* children weight */
unsigned int share; /* disk bandwidth share of the queue */
};
enum tg_state_flags {
THROTL_TG_PENDING = 1 << 0, /* on parent's pending tree */
THROTL_TG_WAS_EMPTY = 1 << 1, /* bio_lists[] became non-empty */
THROTL_TG_ONLINE = 1 << 2, /* tg online */
};
#define rb_entry_tg(node) rb_entry((node), struct throtl_grp, rb_node)
struct throtl_io_cost {
/* bytes per second rate limits */
uint64_t bps[2];
/* IOPS limits */
unsigned int iops[2];
/* Number of bytes disptached in current slice */
uint64_t bytes_disp[2];
/* Number of bio's dispatched in current slice */
unsigned int io_disp[2];
};
struct throtl_grp {
/* must be the first member */
struct blkg_policy_data pd;
/* active throtl group service_queue member */
struct rb_node rb_node;
/* throtl_data this group belongs to */
struct throtl_data *td;
/* this group's service queue */
struct throtl_service_queue service_queue;
/*
* qnode_on_self is used when bios are directly queued to this
* throtl_grp so that local bios compete fairly with bios
* dispatched from children. qnode_on_parent is used when bios are
* dispatched from this throtl_grp into its parent and will compete
* with the sibling qnode_on_parents and the parent's
* qnode_on_self.
*/
struct throtl_qnode qnode_on_self[2];
struct throtl_qnode qnode_on_parent[2];
/*
* Dispatch time in jiffies. This is the estimated time when group
* will unthrottle and is ready to dispatch more bio. It is used as
* key to sort active groups in service tree.
*/
unsigned long disptime;
unsigned int flags;
unsigned int type;
/* are there any throtl rules between this group and td? */
bool has_rules[2];
struct throtl_io_cost io_cost;
/* When did we start a new slice */
unsigned long slice_start[2];
unsigned long slice_end[2];
unsigned long intime;
atomic_t inflights[2];
struct bio_list bios;
struct list_head node;
wait_queue_head_t wait;
bool expired;
int quantum;
atomic_t delay_cnt;
atomic_t max_delay;
};
enum run_mode {
MODE_NONE = 0,
MODE_THROTTLE = 1, /* bandwidth/iops based throttle */
/* below are weight based */
MODE_WEIGHT_BANDWIDTH = 2,
MODE_WEIGHT_IOPS = 3,
MAX_MODE = 4,
};
static char *run_mode_name[MAX_MODE] = {
[MODE_NONE] = "none",
[MODE_THROTTLE] = "throttle",
[MODE_WEIGHT_BANDWIDTH] = "weight_bw",
[MODE_WEIGHT_IOPS] = "weight_iops",
};
unsigned int THROTL_IDLE_INTERVAL = 40; /* ms */
struct throtl_data
{
/* service tree for active throtl groups */
struct throtl_service_queue service_queue;
struct request_queue *queue;
/* Total Number of queued bios on READ and WRITE lists */
unsigned int nr_queued[2];
/* Work for dispatching throttled bios */
struct work_struct dispatch_work;
uint64_t bw_slice;
unsigned int iops_slice;
int quantum;
enum run_mode mode;
int max_inflights;
atomic_t inflights;
struct list_head active;
struct list_head expired;
wait_queue_head_t waitq;
struct timer_list rescue_timer;
struct timer_list idle_timer;
};
static bool td_weight_based(struct throtl_data *td)
{
return td->mode > MODE_THROTTLE;
}
static void throtl_pending_timer_fn(unsigned long arg);
static inline struct throtl_grp *pd_to_tg(struct blkg_policy_data *pd)
{
/*lint -save -e826*/
return pd ? container_of(pd, struct throtl_grp, pd) : NULL;
/*lint -restore*/
}
static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg)
{
return pd_to_tg(blkg_to_pd(blkg, &blkcg_policy_throtl));
}
static inline struct blkcg_gq *tg_to_blkg(struct throtl_grp *tg)
{
return pd_to_blkg(&tg->pd);
}
/**
* sq_to_tg - return the throl_grp the specified service queue belongs to
* @sq: the throtl_service_queue of interest
*
* Return the throtl_grp @sq belongs to. If @sq is the top-level one
* embedded in throtl_data, %NULL is returned.
*/
static struct throtl_grp *sq_to_tg(struct throtl_service_queue *sq)
{
if (sq && sq->parent_sq)
/*lint -save -e826*/
return container_of(sq, struct throtl_grp, service_queue);
/*lint -restore*/
else
return NULL;
}
/**
* sq_to_td - return throtl_data the specified service queue belongs to
* @sq: the throtl_service_queue of interest
*
* A service_queue can be embeded in either a throtl_grp or throtl_data.
* Determine the associated throtl_data accordingly and return it.
*/
static struct throtl_data *sq_to_td(struct throtl_service_queue *sq)
{
struct throtl_grp *tg = sq_to_tg(sq);
if (tg)
return tg->td;
else
/*lint -save -e826*/
return container_of(sq, struct throtl_data, service_queue);
/*lint -restore*/
}
static inline int tg_data_index(struct throtl_grp *tg, bool rw)
{
if (td_weight_based(tg->td))
return 0;
return rw;
}
static inline uint64_t queue_bandwidth_slice(struct throtl_data *td)
{
return td->bw_slice;
}
static inline unsigned int queue_iops_slice(struct throtl_data *td)
{
return td->iops_slice;
}
/**
* throtl_log - log debug message via blktrace
* @sq: the service_queue being reported
* @fmt: printf format string
* @args: printf args
*
* The messages are prefixed with "throtl BLKG_NAME" if @sq belongs to a
* throtl_grp; otherwise, just "throtl".
*/
#define throtl_log(sq, fmt, args...) do { \
struct throtl_grp *__tg = sq_to_tg((sq)); \
struct throtl_data *__td = sq_to_td((sq)); \
\
(void)__td; \
if (likely(!blk_trace_note_message_enabled(__td->queue))) \
break; \
if ((__tg)) { \
char __pbuf[128]; \
\
blkg_path(tg_to_blkg(__tg), __pbuf, sizeof(__pbuf)); \
blk_add_trace_msg(__td->queue, "throtl %s " fmt, __pbuf, ##args); \
} else { \
blk_add_trace_msg(__td->queue, "throtl " fmt, ##args); \
} \
} while (0)
static void throtl_qnode_init(struct throtl_qnode *qn, struct throtl_grp *tg)
{
INIT_LIST_HEAD(&qn->node);
bio_list_init(&qn->bios);
qn->tg = tg;
}
/**
* throtl_qnode_add_bio - add a bio to a throtl_qnode and activate it
* @bio: bio being added
* @qn: qnode to add bio to
* @queued: the service_queue->queued[] list @qn belongs to
*
* Add @bio to @qn and put @qn on @queued if it's not already on.
* @qn->tg's reference count is bumped when @qn is activated. See the
* comment on top of throtl_qnode definition for details.
*/
static void throtl_qnode_add_bio(struct bio *bio, struct throtl_qnode *qn,
struct list_head *queued)
{
bio_list_add(&qn->bios, bio);
if (list_empty(&qn->node)) {
list_add_tail(&qn->node, queued);
blkg_get(tg_to_blkg(qn->tg));
}
}
/**
* throtl_peek_queued - peek the first bio on a qnode list
* @queued: the qnode list to peek
*/
static struct bio *throtl_peek_queued(struct list_head *queued)
{
/*lint -save -e727 -e730 -e826*/
struct throtl_qnode *qn = list_first_entry(queued, struct throtl_qnode, node);
struct bio *bio;
if (list_empty(queued))
return NULL;
bio = bio_list_peek(&qn->bios);
WARN_ON_ONCE(!bio);
/*lint -restore*/
return bio;
}
/**
* throtl_pop_queued - pop the first bio form a qnode list
* @queued: the qnode list to pop a bio from
* @tg_to_put: optional out argument for throtl_grp to put
*
* Pop the first bio from the qnode list @queued. After popping, the first
* qnode is removed from @queued if empty or moved to the end of @queued so
* that the popping order is round-robin.
*
* When the first qnode is removed, its associated throtl_grp should be put
* too. If @tg_to_put is NULL, this function automatically puts it;
* otherwise, *@tg_to_put is set to the throtl_grp to put and the caller is
* responsible for putting it.
*/
static struct bio *throtl_pop_queued(struct list_head *queued,
struct throtl_grp **tg_to_put)
{
/*lint -save -e826*/
struct throtl_qnode *qn = list_first_entry(queued, struct throtl_qnode, node);
/*lint -restore*/
struct bio *bio;
if (list_empty(queued))
return NULL;
bio = bio_list_pop(&qn->bios);
/*lint -save -e727 -e730*/
WARN_ON_ONCE(!bio);
/*lint -restore*/
if (bio_list_empty(&qn->bios)) {
list_del_init(&qn->node);
if (tg_to_put)
*tg_to_put = qn->tg;
else
blkg_put(tg_to_blkg(qn->tg));
} else {
list_move_tail(&qn->node, queued);
}
return bio;
}
/* init a service_queue, assumes the caller zeroed it */
static void throtl_service_queue_init(struct throtl_service_queue *sq)
{
INIT_LIST_HEAD(&sq->queued[0]);
INIT_LIST_HEAD(&sq->queued[1]);
sq->pending_tree = RB_ROOT;
setup_timer(&sq->pending_timer, throtl_pending_timer_fn,
(unsigned long)sq);
}
static inline void io_cost_init(struct throtl_grp *tg)
{
tg->io_cost.bps[READ] = (u64)(s64)-1;
tg->io_cost.bps[WRITE] = (u64)(s64)-1;
tg->io_cost.iops[READ] = (u32)-1;
tg->io_cost.iops[WRITE] = (u32)-1;
}
static struct blkg_policy_data *throtl_pd_alloc(gfp_t gfp, int node)
{
struct throtl_grp *tg;
int rw;
tg = kzalloc_node(sizeof(*tg), gfp, node);
if (!tg)
return NULL;
throtl_service_queue_init(&tg->service_queue);
/*lint -save -e574 -e737 -e1058*/
for (rw = READ; rw <= WRITE; rw++) {
throtl_qnode_init(&tg->qnode_on_self[rw], tg);
throtl_qnode_init(&tg->qnode_on_parent[rw], tg);
}
tg->intime = 0;
atomic_set(&tg->inflights[0], 0);
atomic_set(&tg->inflights[1], 0);
bio_list_init(&tg->bios);
INIT_LIST_HEAD(&tg->node);
init_waitqueue_head(&tg->wait);
tg->expired = false;
/*lint -restore*/
RB_CLEAR_NODE(&tg->rb_node);
io_cost_init(tg);
return &tg->pd;
}
static void throtl_pd_init(struct blkg_policy_data *pd)
{
struct throtl_grp *tg = pd_to_tg(pd);
struct blkcg_gq *blkg = tg_to_blkg(tg);
struct throtl_data *td = blkg->q->td;
struct throtl_service_queue *sq = &tg->service_queue;
/*
* If on the default hierarchy, we switch to properly hierarchical
* behavior where limits on a given throtl_grp are applied to the
* whole subtree rather than just the group itself. e.g. If 16M
* read_bps limit is set on the root group, the whole system can't
* exceed 16M for the device.
*
* If not on the default hierarchy, the broken flat hierarchy
* behavior is retained where all throtl_grps are treated as if
* they're all separate root groups right below throtl_data.
* Limits of a group don't interact with limits of other groups
* regardless of the position of the group in the hierarchy.
*/
sq->parent_sq = &td->service_queue;
if (cgroup_subsys_on_dfl(io_cgrp_subsys) && blkg->parent)
sq->parent_sq = &blkg_to_tg(blkg->parent)->service_queue;
tg->td = td;
}
static inline bool io_cost_has_limit(struct throtl_grp *tg, int index)
{
return tg->io_cost.bps[index] != (u64)(s64)-1 ||
tg->io_cost.iops[index] != (u32)-1;
}
/*
* Set has_rules[] if @tg or any of its parents have limits configured.
* This doesn't require walking up to the top of the hierarchy as the
* parent's has_rules[] is guaranteed to be correct.
*/
static void tg_update_has_rules(struct throtl_grp *tg)
{
struct throtl_grp *parent_tg = sq_to_tg(tg->service_queue.parent_sq);
int i;
/*lint -save -e574 -e737*/
for (i = READ; i <= WRITE; i++)
/*lint -restore*/
tg->has_rules[i] = (parent_tg && parent_tg->has_rules[i]) ||
(io_cost_has_limit(tg, i)) ||
(td_weight_based(tg->td));
}
static void tg_update_perf(struct throtl_grp *tg)
{
struct throtl_service_queue *sq;
u64 new_bps;
unsigned int new_iops;
sq = &tg->service_queue;
if (!sq->parent_sq)
return;
if (tg->td->mode == MODE_WEIGHT_BANDWIDTH) {
new_bps = max_t(uint64_t,
(queue_bandwidth_slice(tg->td) * sq->share) >> SHARE_SHIFT,
1024);
tg->io_cost.bps[0] = new_bps;
tg->io_cost.iops[0] = (u32)-1;
} else if (tg->td->mode == MODE_WEIGHT_IOPS) {
new_iops = max_t(unsigned int,
(queue_iops_slice(tg->td) * sq->share) >> SHARE_SHIFT,
1);
tg->io_cost.iops[0] = new_iops;
tg->io_cost.bps[0] = (u64)(s64)-1;
}
/*lint -save -e737*/
tg->quantum = max_t(int,
(throtl_quantum * sq->share) >> SHARE_SHIFT,
1);
/*lint -restore*/
}
/* update share of tg's siblings */
static void tg_update_share(struct throtl_data *td, struct throtl_grp *tg)
{
struct cgroup_subsys_state *pos_css;
struct blkcg_gq *blkg, *parent_blkg;
struct throtl_grp *child;
if (!tg || !tg->service_queue.parent_sq ||
!tg->service_queue.parent_sq->parent_sq)
parent_blkg = td->queue->root_blkg;
else
parent_blkg = tg_to_blkg(sq_to_tg(tg->service_queue.parent_sq));
/*lint -save -e747 -e820*/
blkg_for_each_descendant_pre(blkg, pos_css, parent_blkg) {
/*lint -restore*/
struct throtl_service_queue *sq;
child = blkg_to_tg(blkg);
sq = &child->service_queue;
if (!sq->parent_sq || !(child->flags & THROTL_TG_ONLINE))
continue;
sq->share = max_t(unsigned int,
sq->parent_sq->share * blkg->weight /
sq->parent_sq->children_weight,
1);
tg_update_perf(child);
}
}
static void throtl_pd_online(struct blkg_policy_data *pd)
{
struct throtl_grp *tg = pd_to_tg(pd);
struct throtl_service_queue *parent_sq;
struct blkcg_gq *blkg = tg_to_blkg(tg);
/*
* We don't want new groups to escape the limits of its ancestors.
* Update has_rules[] after a new group is brought online.
*/
tg_update_has_rules(tg);
tg->flags |= THROTL_TG_ONLINE;
parent_sq = tg->service_queue.parent_sq;
if (parent_sq)
parent_sq->children_weight += blkg->weight;
tg_update_share(tg->td, tg);
}
static void throtl_pd_offline(struct blkg_policy_data *pd)
{
struct throtl_grp *tg = pd_to_tg(pd);
struct throtl_service_queue *sq = &tg->service_queue;
struct blkcg_gq *blkg = tg_to_blkg(tg);
if (!(tg->flags & THROTL_TG_ONLINE))
return;
if (bio_list_empty(&tg->bios) && !list_empty(&tg->node)) {
list_del_init(&tg->node);
blkg_put(blkg);
}
tg->flags &= ~THROTL_TG_ONLINE;
if (sq->parent_sq)
sq->parent_sq->children_weight -= blkg->weight;
rcu_read_lock();
tg_update_share(tg->td, tg);
rcu_read_unlock();
}
static void throtl_pd_free(struct blkg_policy_data *pd)
{
struct throtl_grp *tg = pd_to_tg(pd);
del_timer_sync(&tg->service_queue.pending_timer);
kfree(tg);
}
static struct throtl_grp *
throtl_rb_first(struct throtl_service_queue *parent_sq)
{
/* Service tree is empty */
if (!parent_sq->nr_pending)
return NULL;
if (!parent_sq->first_pending)
parent_sq->first_pending = rb_first(&parent_sq->pending_tree);
if (parent_sq->first_pending)
/*lint -save -e826*/
return rb_entry_tg(parent_sq->first_pending);
/*lint -restore*/
return NULL;
}
static void rb_erase_init(struct rb_node *n, struct rb_root *root)
{
rb_erase(n, root);
RB_CLEAR_NODE(n);
}
static void throtl_rb_erase(struct rb_node *n,
struct throtl_service_queue *parent_sq)
{
if (parent_sq->first_pending == n)
parent_sq->first_pending = NULL;
rb_erase_init(n, &parent_sq->pending_tree);
--parent_sq->nr_pending;
}
static void update_min_dispatch_time(struct throtl_service_queue *parent_sq)
{
struct throtl_grp *tg;
tg = throtl_rb_first(parent_sq);
if (!tg)
return;
parent_sq->first_pending_disptime = tg->disptime;
}
static void tg_service_queue_add(struct throtl_grp *tg)
{
struct throtl_service_queue *parent_sq = tg->service_queue.parent_sq;
struct rb_node **node = &parent_sq->pending_tree.rb_node;
struct rb_node *parent = NULL;
struct throtl_grp *__tg;
unsigned long key = tg->disptime;
int left = 1;
while (*node != NULL) {
parent = *node;
/*lint -save -e550 -e774 -e826*/
__tg = rb_entry_tg(parent);
if (time_before(key, __tg->disptime))
node = &parent->rb_left;
else {
node = &parent->rb_right;
left = 0;
}
/*lint -restore*/
}
if (left)
parent_sq->first_pending = &tg->rb_node;
rb_link_node(&tg->rb_node, parent, node);
rb_insert_color(&tg->rb_node, &parent_sq->pending_tree);
}
static void __throtl_enqueue_tg(struct throtl_grp *tg)
{
tg_service_queue_add(tg);
tg->flags |= THROTL_TG_PENDING;
tg->service_queue.parent_sq->nr_pending++;
}
static void throtl_enqueue_tg(struct throtl_grp *tg)
{
if (!(tg->flags & THROTL_TG_PENDING))
__throtl_enqueue_tg(tg);
}
static void __throtl_dequeue_tg(struct throtl_grp *tg)
{
throtl_rb_erase(&tg->rb_node, tg->service_queue.parent_sq);
tg->flags &= ~THROTL_TG_PENDING;
}
static void throtl_dequeue_tg(struct throtl_grp *tg)
{
if (tg->flags & THROTL_TG_PENDING)
__throtl_dequeue_tg(tg);
}
/* Call with queue lock held */
static void throtl_schedule_pending_timer(struct throtl_service_queue *sq,
unsigned long expires)
{
unsigned long max_expire = jiffies + 8 * throtl_slice;
/*
* Since we are adjusting the throttle limit dynamically, the sleep
* time calculated according to previous limit might be invalid. It's
* possible the cgroup sleep time is very long and no other cgroups
* have IO running so notify the limit changes. Make sure the cgroup
* doesn't sleep too long to avoid the missed notification.
*/
if (time_after(expires, max_expire))
expires = max_expire;
mod_timer(&sq->pending_timer, expires);
/*lint -save -e747*/
throtl_log(sq, "schedule timer. delay=%lu jiffies=%lu",
expires - jiffies, jiffies);
/*lint -restore*/
}
/**
* throtl_schedule_next_dispatch - schedule the next dispatch cycle
* @sq: the service_queue to schedule dispatch for
* @force: force scheduling
*
* Arm @sq->pending_timer so that the next dispatch cycle starts on the
* dispatch time of the first pending child. Returns %true if either timer
* is armed or there's no pending child left. %false if the current
* dispatch window is still open and the caller should continue
* dispatching.
*
* If @force is %true, the dispatch timer is always scheduled and this
* function is guaranteed to return %true. This is to be used when the
* caller can't dispatch itself and needs to invoke pending_timer
* unconditionally. Note that forced scheduling is likely to induce short
* delay before dispatch starts even if @sq->first_pending_disptime is not
* in the future and thus shouldn't be used in hot paths.
*/
static bool throtl_schedule_next_dispatch(struct throtl_service_queue *sq,
bool force)
{
/* any pending children left? */
if (!sq->nr_pending)
return true;
update_min_dispatch_time(sq);
/* is the next dispatch time in the future? */
/*lint -save -e550 -e774*/
if (force || time_after(sq->first_pending_disptime, jiffies)) {
throtl_schedule_pending_timer(sq, sq->first_pending_disptime);
return true;
}
/*lint -restore*/
/* tell the caller to continue dispatching */
return false;
}
static inline void io_cost_start_new_slice(struct throtl_grp *tg,
bool rw)
{
int index = tg_data_index(tg, rw);
tg->io_cost.bytes_disp[index] = 0;
tg->io_cost.io_disp[index] = 0;
}
static inline void throtl_start_new_slice_with_credit(struct throtl_grp *tg,
bool rw, unsigned long start)
{
int index = tg_data_index(tg, rw);
io_cost_start_new_slice(tg, rw);
/*
* Previous slice has expired. We must have trimmed it after last
* bio dispatch. That means since start of last slice, we never used
* that bandwidth. Do try to make use of that bandwidth while giving
* credit.
*/
/*lint -save -e550 -e747 -e774 -e1072*/
if (time_after_eq(start, tg->slice_start[index]))
tg->slice_start[index] = start;
tg->slice_end[index] = jiffies + throtl_slice;
throtl_log(&tg->service_queue,
"[%c] new slice with credit start=%lu end=%lu jiffies=%lu",
rw == READ ? 'R' : 'W', tg->slice_start[index],
tg->slice_end[index], jiffies);
/*lint -restore*/
}
static inline void throtl_start_new_slice(struct throtl_grp *tg, bool rw)
{
int index = tg_data_index(tg, rw);
io_cost_start_new_slice(tg, rw);
tg->slice_start[index] = jiffies;
tg->slice_end[index] = jiffies + throtl_slice;
/*lint -save -e747*/
throtl_log(&tg->service_queue,
"[%c] new slice start=%lu end=%lu jiffies=%lu",
rw == READ ? 'R' : 'W', tg->slice_start[index],
tg->slice_end[index], jiffies);
/*lint -restore*/
}
static inline void throtl_set_slice_end(struct throtl_grp *tg, bool rw,
unsigned long jiffy_end)
{
int index = tg_data_index(tg, rw);
tg->slice_end[index] = roundup(jiffy_end, throtl_slice);
}
static inline void throtl_extend_slice(struct throtl_grp *tg, bool rw,
unsigned long jiffy_end)
{
int index = tg_data_index(tg, rw);
tg->slice_end[index] = roundup(jiffy_end, throtl_slice);
/*lint -save -e747*/
throtl_log(&tg->service_queue,
"[%c] extend slice start=%lu end=%lu jiffies=%lu",
rw == READ ? 'R' : 'W', tg->slice_start[index],
tg->slice_end[index], jiffies);
/*lint -restore*/
}
/* Determine if previously allocated or extended slice is complete or not */
static bool throtl_slice_used(struct throtl_grp *tg, bool rw)
{
int index = tg_data_index(tg, rw);
/*lint -save -e550 -e774 -e1072*/
if (time_in_range(jiffies, tg->slice_start[index], tg->slice_end[index]))
/*lint -restore*/
return false;
return 1;
}
static inline bool io_cost_trim_slice(struct throtl_grp *tg, bool rw,
unsigned long nr_slices)
{
unsigned long io_trim;
u64 bytes_trim, tmp;
int index = tg_data_index(tg, rw);
tmp = tg->io_cost.bps[index] * throtl_slice * nr_slices;
do_div(tmp, HZ);
bytes_trim = tmp;
io_trim = (tg->io_cost.iops[index] * throtl_slice * nr_slices) / HZ;
if (!bytes_trim && !io_trim)
return false;
if (tg->io_cost.bytes_disp[index] >= bytes_trim)
tg->io_cost.bytes_disp[index] -= bytes_trim;
else
tg->io_cost.bytes_disp[index] = 0;
if (tg->io_cost.io_disp[index] >= io_trim)
tg->io_cost.io_disp[index] -= (unsigned int)io_trim;
else
tg->io_cost.io_disp[index] = 0;
return true;
}
/* Trim the used slices and adjust slice start accordingly */
static inline void throtl_trim_slice(struct throtl_grp *tg, bool rw)
{
unsigned long nr_slices, time_elapsed;
int index = tg_data_index(tg, rw);
/*lint -save -e550 -e730 -e1072*/
BUG_ON(time_before(tg->slice_end[index], tg->slice_start[index]));
/*lint -restore*/
/*
* If bps are unlimited (-1), then time slice don't get
* renewed. Don't try to trim the slice if slice is used. A new
* slice will start when appropriate.
*/
if (throtl_slice_used(tg, rw))
return;
/*
* A bio has been dispatched. Also adjust slice_end. It might happen
* that initially cgroup limit was very low resulting in high
* slice_end, but later limit was bumped up and bio was dispached
* sooner, then we need to reduce slice_end. A high bogus slice_end
* is bad because it does not allow new slice to start.
*/
throtl_set_slice_end(tg, rw, jiffies + throtl_slice);
time_elapsed = jiffies - tg->slice_start[index];
nr_slices = time_elapsed / throtl_slice;
if (!nr_slices)
return;
if (!io_cost_trim_slice(tg, rw, nr_slices))
return;
tg->slice_start[index] += nr_slices * throtl_slice;
/*lint -save -e747*/
throtl_log(&tg->service_queue,
"[%c] trim slice nr=%lu start=%lu end=%lu jiffies=%lu",
rw == READ ? 'R' : 'W', nr_slices,
tg->slice_start[index], tg->slice_end[index], jiffies);
/*lint -restore*/
}
static bool io_cost_with_in_iops_limit(struct throtl_grp *tg, struct bio *bio,
unsigned long *wait)
{
bool rw = bio_data_dir(bio);
int index = tg_data_index(tg, rw);
unsigned int io_allowed;
unsigned long jiffy_elapsed, jiffy_wait, jiffy_elapsed_rnd;
u64 tmp;
jiffy_elapsed = jiffy_elapsed_rnd = jiffies - tg->slice_start[index];
/* Slice has just started. Consider one slice interval */
if (!jiffy_elapsed)
jiffy_elapsed_rnd = throtl_slice;
jiffy_elapsed_rnd = roundup(jiffy_elapsed_rnd, throtl_slice);
/*
* jiffy_elapsed_rnd should not be a big value as minimum iops can be
* 1 then at max jiffy elapsed should be equivalent of 1 second as we
* will allow dispatch after 1 second and after that slice should
* have been trimmed.
*/
tmp = (u64)tg->io_cost.iops[index] * jiffy_elapsed_rnd;
do_div(tmp, HZ);
if (tmp > UINT_MAX)
io_allowed = UINT_MAX;
else
/*lint -save -e712*/
io_allowed = tmp;
/*lint -restore*/
if (tg->io_cost.io_disp[index] + 1 <= io_allowed) {
if (wait)
*wait = 0;
return true;
}
/* Calc approx time to dispatch */
/*lint -save -e647*/
jiffy_wait = ((tg->io_cost.io_disp[index] + 1) * HZ) /
tg->io_cost.iops[index] + 1;
/*lint -restore*/
if (jiffy_wait > jiffy_elapsed)
jiffy_wait = jiffy_wait - jiffy_elapsed;
else
jiffy_wait = 1;
if (wait)
*wait = jiffy_wait;
return false;
}
static bool io_cost_with_in_bps_limit(struct throtl_grp *tg, struct bio *bio,
unsigned long *wait)
{
bool rw = bio_data_dir(bio);
int index = tg_data_index(tg, rw);
u64 bytes_allowed, extra_bytes, tmp;
unsigned long jiffy_elapsed, jiffy_wait, jiffy_elapsed_rnd;
jiffy_elapsed = jiffy_elapsed_rnd = jiffies - tg->slice_start[index];
/* Slice has just started. Consider one slice interval */
if (!jiffy_elapsed)
jiffy_elapsed_rnd = throtl_slice;
jiffy_elapsed_rnd = roundup(jiffy_elapsed_rnd, throtl_slice);
tmp = tg->io_cost.bps[index] * jiffy_elapsed_rnd;
do_div(tmp, HZ);
bytes_allowed = tmp;
if (tg->io_cost.bytes_disp[index] + bio->bi_iter.bi_size <=
bytes_allowed) {
if (wait)
*wait = 0;
return true;
}
/* Calc approx time to dispatch */
extra_bytes = tg->io_cost.bytes_disp[index] +
bio->bi_iter.bi_size - bytes_allowed;
jiffy_wait = div64_u64(extra_bytes * HZ, tg->io_cost.bps[index]);
if (!jiffy_wait)
jiffy_wait = 1;
/*
* This wait time is without taking into consideration the rounding
* up we did. Add that time also.
*/
jiffy_wait = jiffy_wait + (jiffy_elapsed_rnd - jiffy_elapsed);
if (wait)
*wait = jiffy_wait;
return false;
}
static bool io_cost_with_in_limit(struct throtl_grp *tg, struct bio *bio,
unsigned long *wait)
{
unsigned long bps_wait = 0, iops_wait = 0;
if (io_cost_with_in_bps_limit(tg, bio, &bps_wait) &&
io_cost_with_in_iops_limit(tg, bio, &iops_wait)) {
*wait = 0;
return true;
}
*wait = max(bps_wait, iops_wait);
return false;
}
/*
* Returns whether one can dispatch a bio or not. Also returns approx number
* of jiffies to wait before this bio is with-in IO rate and can be dispatched
*/
static bool tg_may_dispatch(struct throtl_grp *tg, struct bio *bio,
unsigned long *wait)
{
bool rw = bio_data_dir(bio);
int index = tg_data_index(tg, rw);
unsigned long max_wait = 0;
/*
* Currently whole state machine of group depends on first bio
* queued in the group bio list. So one should not be calling
* this function with a different bio if there are other bios
* queued.
*/
/*lint -save -e730*/
BUG_ON(tg->service_queue.nr_queued[index] &&
bio != throtl_peek_queued(&tg->service_queue.queued[index]));
/*lint -restore*/
/* If tg->bps = -1, then BW is unlimited */
if (!io_cost_has_limit(tg, index)) {
if (wait)
*wait = 0;
return true;
}
/*
* If previous slice expired, start a new one otherwise renew/extend
* existing slice to make sure it is at least throtl_slice interval
* long since now. New slice is started only for empty throttle group.
* If there is queued bio, that means there should be an active
* slice and it should be extended instead.
*/
if (throtl_slice_used(tg, rw) && !(tg->service_queue.nr_queued[rw]))
throtl_start_new_slice(tg, rw);
else {
/*lint -save -e550 -e774 -e1072*/
if (time_before(tg->slice_end[index], jiffies + throtl_slice))
throtl_extend_slice(tg, rw, jiffies + throtl_slice);
/*lint -restore*/
}
if (io_cost_with_in_limit(tg, bio, &max_wait)) {
if (wait)
*wait = 0;
return true;
}
if (wait)
*wait = max_wait;
/*lint -save -e550 -e774 -e1072*/
if (time_before(tg->slice_end[index], jiffies + max_wait))
throtl_extend_slice(tg, rw, jiffies + max_wait);
/*lint -restore*/
return false;
}
static inline void io_cost_charge_bio(struct throtl_grp *tg, struct bio *bio)
{
bool rw = bio_data_dir(bio);
int index = tg_data_index(tg, rw);
tg->io_cost.bytes_disp[index] += bio->bi_iter.bi_size;
tg->io_cost.io_disp[index]++;
}
static void throtl_charge_bio(struct throtl_grp *tg, struct bio *bio)
{
io_cost_charge_bio(tg, bio);
/*
* REQ_THROTTLED is used to prevent the same bio to be throttled
* more than once as a throttled bio will go through blk-throtl the
* second time when it eventually gets issued. Set it when a bio
* is being charged to a tg.
*/
/*lint -save -e712 -e747*/
if (!(bio->bi_opf & REQ_THROTTLED))
bio->bi_opf |= REQ_THROTTLED;
/*lint -restore*/
}
/**
* throtl_add_bio_tg - add a bio to the specified throtl_grp
* @bio: bio to add
* @qn: qnode to use
* @tg: the target throtl_grp
*
* Add @bio to @tg's service_queue using @qn. If @qn is not specified,
* tg->qnode_on_self[] is used.
*/
static void throtl_add_bio_tg(struct bio *bio, struct throtl_qnode *qn,
struct throtl_grp *tg)
{
struct throtl_service_queue *sq = &tg->service_queue;
bool rw = bio_data_dir(bio);
int index = tg_data_index(tg, rw);
if (!qn)
qn = &tg->qnode_on_self[index];
/*
* If @tg doesn't currently have any bios queued in the same
* direction, queueing @bio can change when @tg should be
* dispatched. Mark that @tg was empty. This is automatically
* cleaered on the next tg_update_disptime().
*/
if (!sq->nr_queued[index])
tg->flags |= THROTL_TG_WAS_EMPTY;
throtl_qnode_add_bio(bio, qn, &sq->queued[index]);
sq->nr_queued[index]++;
throtl_enqueue_tg(tg);
}
static void tg_update_disptime(struct throtl_grp *tg)
{
/*lint -save -e570 -e820 -e838*/
struct throtl_service_queue *sq = &tg->service_queue;
unsigned long read_wait = -1, write_wait = -1, min_wait = -1, disptime;
struct bio *bio;
if ((bio = throtl_peek_queued(&sq->queued[READ])))
tg_may_dispatch(tg, bio, &read_wait);
if ((bio = throtl_peek_queued(&sq->queued[WRITE])))
tg_may_dispatch(tg, bio, &write_wait);
min_wait = min(read_wait, write_wait);
disptime = jiffies + min_wait;
/*lint -restore*/
/* Update dispatch time */
throtl_dequeue_tg(tg);
tg->disptime = disptime;
throtl_enqueue_tg(tg);
/* see throtl_add_bio_tg() */
tg->flags &= ~THROTL_TG_WAS_EMPTY;
}
static void start_parent_slice_with_credit(struct throtl_grp *child_tg,
struct throtl_grp *parent_tg, bool rw)
{
int index = tg_data_index(child_tg, rw);
if (throtl_slice_used(parent_tg, rw)) {
throtl_start_new_slice_with_credit(parent_tg, rw,
child_tg->slice_start[index]);
}
}
static void tg_dispatch_one_bio(struct throtl_grp *tg, bool rw)
{
struct throtl_service_queue *sq = &tg->service_queue;
struct throtl_service_queue *parent_sq = sq->parent_sq;
struct throtl_grp *parent_tg = sq_to_tg(parent_sq);
struct throtl_grp *tg_to_put = NULL;
struct bio *bio;
int index = tg_data_index(tg, rw);
/*
* @bio is being transferred from @tg to @parent_sq. Popping a bio
* from @tg may put its reference and @parent_sq might end up
* getting released prematurely. Remember the tg to put and put it
* after @bio is transferred to @parent_sq.
*/
bio = throtl_pop_queued(&sq->queued[index], &tg_to_put);
sq->nr_queued[index]--;
throtl_charge_bio(tg, bio);
/*
* If our parent is another tg, we just need to transfer @bio to
* the parent using throtl_add_bio_tg(). If our parent is
* @td->service_queue, @bio is ready to be issued. Put it on its
* bio_lists[] and decrease total number queued. The caller is
* responsible for issuing these bios.
*/
if (parent_tg) {
throtl_add_bio_tg(bio, &tg->qnode_on_parent[index], parent_tg);
start_parent_slice_with_credit(tg, parent_tg, rw);
} else {
throtl_qnode_add_bio(bio, &tg->qnode_on_parent[index],
&parent_sq->queued[index]);
/*lint -save -e730 -e775*/
BUG_ON(tg->td->nr_queued[index] <= 0);
/*lint -restore*/
tg->td->nr_queued[index]--;
}
throtl_trim_slice(tg, rw);
if (tg_to_put)
blkg_put(tg_to_blkg(tg_to_put));
}
static int throtl_dispatch_tg(struct throtl_grp *tg)
{
/*lint -save -e713 -e730 -e732 -e737 -e747 -e820*/
struct throtl_service_queue *sq = &tg->service_queue;
unsigned int nr_reads = 0, nr_writes = 0;
unsigned int max_nr_reads = throtl_grp_quantum*3/4;
unsigned int max_nr_writes = throtl_grp_quantum - max_nr_reads;
struct bio *bio;
/* Try to dispatch 75% READS and 25% WRITES */
while ((bio = throtl_peek_queued(&sq->queued[READ])) &&
tg_may_dispatch(tg, bio, NULL)) {
tg_dispatch_one_bio(tg, bio_data_dir(bio));
nr_reads++;
if (nr_reads >= max_nr_reads)
break;
}
while ((bio = throtl_peek_queued(&sq->queued[WRITE])) &&
tg_may_dispatch(tg, bio, NULL)) {
tg_dispatch_one_bio(tg, bio_data_dir(bio));
nr_writes++;
if (nr_writes >= max_nr_writes)
break;
}
return nr_reads + nr_writes;
/*lint -restore*/
}
static int throtl_select_dispatch(struct throtl_service_queue *parent_sq)
{
unsigned int nr_disp = 0;
/*lint -save -e550 -e574 -e713 -e737 -e774*/
while (1) {
struct throtl_grp *tg = throtl_rb_first(parent_sq);
struct throtl_service_queue *sq = &tg->service_queue;
if (!tg)
break;
if (time_before(jiffies, tg->disptime))
break;
throtl_dequeue_tg(tg);
nr_disp += throtl_dispatch_tg(tg);
if (sq->nr_queued[0] || sq->nr_queued[1])
tg_update_disptime(tg);
if (nr_disp >= throtl_quantum)
break;
}
return nr_disp;
/*lint -restore*/
}
/**
* throtl_pending_timer_fn - timer function for service_queue->pending_timer
* @arg: the throtl_service_queue being serviced
*
* This timer is armed when a child throtl_grp with active bio's become
* pending and queued on the service_queue's pending_tree and expires when
* the first child throtl_grp should be dispatched. This function
* dispatches bio's from the children throtl_grps to the parent
* service_queue.
*
* If the parent's parent is another throtl_grp, dispatching is propagated
* by either arming its pending_timer or repeating dispatch directly. If
* the top-level service_tree is reached, throtl_data->dispatch_work is
* kicked so that the ready bio's are issued.
*/
static void throtl_pending_timer_fn(unsigned long arg)
{
struct throtl_service_queue *sq = (void *)arg;
struct throtl_grp *tg = sq_to_tg(sq);
struct throtl_data *td = sq_to_td(sq);
struct request_queue *q = td->queue;
struct throtl_service_queue *parent_sq;
bool dispatched;
int ret;
spin_lock_irq(q->queue_lock);
again:
parent_sq = sq->parent_sq;
dispatched = false;
/*lint -save -e747*/
while (true) {
throtl_log(sq, "dispatch nr_queued=%u read=%u write=%u",
sq->nr_queued[READ] + sq->nr_queued[WRITE],
sq->nr_queued[READ], sq->nr_queued[WRITE]);
ret = throtl_select_dispatch(sq);
if (ret) {
throtl_log(sq, "bios disp=%u", ret);
dispatched = true;
}
if (throtl_schedule_next_dispatch(sq, false))
break;
/* this dispatch windows is still open, relax and repeat */
spin_unlock_irq(q->queue_lock);
cpu_relax();
spin_lock_irq(q->queue_lock);
}
if (!dispatched)
goto out_unlock;
if (parent_sq) {
/* @parent_sq is another throl_grp, propagate dispatch */
if (tg->flags & THROTL_TG_WAS_EMPTY) {
tg_update_disptime(tg);
if (!throtl_schedule_next_dispatch(parent_sq, false)) {
/* window is already open, repeat dispatching */
sq = parent_sq;
tg = sq_to_tg(sq);
goto again;
}
}
} else {
/* reached the top-level, queue issueing */
queue_work(kthrotld_workqueue, &td->dispatch_work);
}
/*lint -restore*/
out_unlock:
spin_unlock_irq(q->queue_lock);
}
/**
* blk_throtl_dispatch_work_fn - work function for throtl_data->dispatch_work
* @work: work item being executed
*
* This function is queued for execution when bio's reach the bio_lists[]
* of throtl_data->service_queue. Those bio's are ready and issued by this
* function.
*/
static void blk_throtl_dispatch_work_fn(struct work_struct *work)
{
/*lint -save -e826*/
struct throtl_data *td = container_of(work, struct throtl_data,
dispatch_work);
/*lint -restore*/
struct throtl_service_queue *td_sq = &td->service_queue;
struct request_queue *q = td->queue;
struct bio_list bio_list_on_stack;
struct bio *bio;
struct blk_plug plug;
int i;
bio_list_init(&bio_list_on_stack);
spin_lock_irq(q->queue_lock);
/*lint -save -e574 -e737 -e820*/
for (i = READ; i <= WRITE; i++)
while ((bio = throtl_pop_queued(&td_sq->queued[i], NULL)))
bio_list_add(&bio_list_on_stack, bio);
spin_unlock_irq(q->queue_lock);
if (!bio_list_empty(&bio_list_on_stack)) {
blk_start_plug(&plug);
while((bio = bio_list_pop(&bio_list_on_stack)))
generic_make_request(bio);
blk_finish_plug(&plug);
}
/*lint -restore*/
}
static bool tg_may_dispatch_weight(struct throtl_grp *tg)
{
if (!io_cost_has_limit(tg, 0))
return true;
if (!(tg->flags & THROTL_TG_ONLINE))
return true;
if (tg->td->mode == MODE_WEIGHT_BANDWIDTH)
return (tg->io_cost.bytes_disp[0] < tg->io_cost.bps[0]);
else if (tg->td->mode == MODE_WEIGHT_IOPS)
return (tg->io_cost.io_disp[0] < tg->io_cost.iops[0]);
else
return true;
}
static void throtl_add_bio_weight_trace(struct bio *bio)
{
bio->bi_throtl_in_queue = jiffies;
#ifdef CONFIG_HUAWEI_IO_TRACING
trace_block_throttle_bio_in(bio);
#endif
}
static void throtl_pop_bio_weight_trace(struct bio *bio, struct throtl_grp *tg)
{
unsigned int delay;
char b[BDEVNAME_SIZE];
delay = jiffies_to_msecs(jiffies - bio->bi_throtl_in_queue);
if (delay > MAX_THROTL_DELAY_TM_MS) {
/*lint -save -e529 -e438 -e1058*/
if ((unsigned int)atomic_read(&tg->max_delay) < delay)
atomic_set(&tg->max_delay, (int)delay);
atomic_inc(&tg->delay_cnt);
/*lint -restore*/
printk("io-latency: [throtl] [%s] sector=%lu size=%u flag=0x%lx delay=%d weight=%d type=%d\n",
bdevname(bio->bi_bdev, b), bio->bi_iter.bi_sector,
bio->bi_iter.bi_size, (unsigned long)bio->bi_opf, delay,
tg_to_blkg(tg)->weight, tg_to_blkg(tg)->blkcg->type);
}
#ifdef CONFIG_HUAWEI_IO_TRACING
trace_block_throttle_bio_out(bio, (long)delay);
#endif
}
static void throtl_add_bio_weight(struct bio *bio, struct throtl_grp *tg)
{
struct throtl_data *td = tg->td;
throtl_add_bio_weight_trace(bio);
#ifdef CONFIG_HISI_BLK_CORE
bio->io_in_count |= HISI_IO_IN_COUNT_WILL_BE_SEND_AGAIN;
#endif
bio_list_add(&tg->bios, bio);
tg->service_queue.nr_queued[0]++;
tg->td->nr_queued[0]++;
tg->intime = jiffies;
#ifdef CONFIG_HUAWEI_IO_TRACING
trace_block_throttle_weight(bio, tg_to_blkg(tg)->weight,
tg->td->nr_queued[0]);
#endif
if (tg->td->nr_queued[0] == 1 && !timer_pending(&tg->td->rescue_timer))
mod_timer(&tg->td->rescue_timer,
jiffies + msecs_to_jiffies(THROTL_IDLE_INTERVAL));
if (!list_empty(&tg->node))
return;
blkg_get(tg_to_blkg(tg));
if (tg_may_dispatch_weight(tg))
list_add_tail(&tg->node, &td->active);
else
list_add_tail(&tg->node, &td->expired);
}
static struct bio *throtl_pop_queued_weight(struct throtl_grp *tg)
{
struct bio *bio;
bio = bio_list_pop(&tg->bios);
/*lint -save -e727 -e730*/
WARN_ON_ONCE(!bio);
/*lint -restore*/
throtl_pop_bio_weight_trace(bio, tg);
return bio;
}
static void blk_throtl_weight_io_done(struct bio *bio)
{
struct throtl_grp *tg;
struct throtl_data *td;
int index = (int)bio_data_dir(bio);
int inflights;
tg = (struct throtl_grp *)bio->bi_throtl_private2;
td = tg->td;
atomic_dec(&tg->inflights[index]);
inflights = atomic_dec_return(&td->inflights);
smp_mb__after_atomic();
if (inflights >= td->quantum)
goto out;
else if (inflights == 0 && waitqueue_active(&td->waitq))
wake_up(&td->waitq);
if (td->nr_queued[0])
queue_work(kthrotld_workqueue, &td->dispatch_work);
out:
bio->bi_throtl_private2 = NULL;
bio->bi_throtl_end_io2 = NULL;
blkg_put(tg_to_blkg(tg));
}
static void tg_dispatch_one_bio_weight(struct throtl_grp *tg,
struct bio_list *bios, bool charge)
{
struct bio *bio;
/*
* @bio is being transferred from @tg to @parent_sq. Popping a bio
* from @tg may put its reference and @parent_sq might end up
* getting released prematurely. Remember the tg to put and put it
* after @bio is transferred to @parent_sq.
*/
bio = throtl_pop_queued_weight(tg);
tg->service_queue.nr_queued[0]--;
/*lint -save -e730*/
if (unlikely(!charge))
/*lint -restore*/
goto skip_charge;
throtl_charge_bio(tg, bio);
if (!bio->bi_throtl_private2) {
int index = (int)bio_data_dir(bio);
atomic_inc(&tg->inflights[index]);
/* Don't reorder td->nr_queued and inflights */
smp_mb__before_atomic();
atomic_inc(&tg->td->inflights);
smp_mb__after_atomic();
bio->bi_throtl_private2 = tg;
/* We might access tg when end io, so grab it */
blkg_get(tg_to_blkg(tg));
bio->bi_throtl_end_io2 = blk_throtl_weight_io_done;
}
skip_charge:
/*lint -save -e730 -e775*/
BUG_ON(tg->td->nr_queued[0] <= 0);
/*lint -restore*/
tg->td->nr_queued[0]--;
bio_list_add(bios, bio);
#ifdef CONFIG_HUAWEI_IO_TRACING
trace_block_throttle_dispatch(bio, tg_to_blkg(tg)->weight);
trace_block_throttle_iocost(tg->io_cost.bps[0],
tg->io_cost.iops[0], tg->io_cost.bytes_disp[0],
tg->io_cost.io_disp[0]);
#endif
}
static int throtl_dispatch_tg_weight(struct throtl_grp *tg,
struct bio_list *bios)
{
int nr = 0;
/*lint -save -e730*/
BUG_ON(!td_weight_based(tg->td));
/*lint -restore*/
while (!bio_list_empty(&tg->bios) &&
tg_may_dispatch_weight(tg)) {
tg_dispatch_one_bio_weight(tg, bios, (bool)true);
nr++;
if (nr >= tg->quantum)
break;
}
if (bio_list_empty(&tg->bios) && !(tg->flags & THROTL_TG_ONLINE)) {
list_del_init(&tg->node);
blkg_put(tg_to_blkg(tg));
} else if (!tg_may_dispatch_weight(tg)) {
list_move_tail(&tg->node, &tg->td->expired);
}
return nr;
}
static void throtl_start_new_slice_weight(struct throtl_data *td)
{
struct throtl_grp *tg;
/*lint -save -e64 -e826*/
list_for_each_entry(tg, &td->expired, node) {
/*lint -restore*/
tg->expired = false;
if (tg->td->mode == MODE_WEIGHT_BANDWIDTH) {
tg->io_cost.bytes_disp[0] = 0;
wake_up_all(&tg->wait);
} else {
wake_up_nr(&tg->wait, tg->io_cost.iops[0]);
tg->io_cost.io_disp[0] = 0;
}
}
list_splice_tail_init(&td->expired, &td->active);
}
static bool should_start_new_slice(struct throtl_data *td)
{
struct throtl_grp *tg;
unsigned long interval;
interval = msecs_to_jiffies(THROTL_IDLE_INTERVAL / 2);
/*lint -save -e64 -e550 -e774 -e826*/
list_for_each_entry(tg, &td->expired, node) {
if (tg_to_blkg(tg)->blkcg->type <= BLK_THROTL_FG)
return true;
}
list_for_each_entry(tg, &td->active, node) {
if (time_before(jiffies, tg->intime + interval))
return false;
}
/*lint -restore*/
return true;
}
static void blk_run_throtl_bios_weight(struct throtl_data *td,
struct throtl_grp *tg,
struct bio_list *bios)
{
struct throtl_grp *curr, *next;
if (tg)
throtl_dispatch_tg_weight(tg, bios);
if (list_empty(&td->active) && bio_list_empty(bios))
goto start_new_slice;
/*lint -save -e64 -e530 -e826*/
list_for_each_entry_safe(curr, next, &td->active, node) {
/*lint -restore*/
if (curr == tg)
continue;
throtl_dispatch_tg_weight(curr, bios);
}
if (!bio_list_empty(bios))
return;
/*
* If no bios in active queues and many bios are pending in
* the device queue, just idle.
*/
start_new_slice:
if (!should_start_new_slice(td)) {
unsigned int interval = THROTL_IDLE_INTERVAL / 2;
mod_timer(&td->idle_timer,
jiffies + msecs_to_jiffies(interval));
return;
}
throtl_start_new_slice_weight(td);
/*lint -save -e64 -e826*/
list_for_each_entry_safe(curr, next, &td->active, node)
throtl_dispatch_tg_weight(curr, bios);
/*lint -restore*/
}
static void blk_throtl_kick_queue_fn(struct work_struct *work)
{
/*lint -save -e826*/
struct throtl_data *td = container_of(work, struct throtl_data,
dispatch_work);
/*lint -restore*/
struct request_queue *q = td->queue;
struct bio_list bio_list_on_stack;
struct bio *bio;
struct blk_plug plug;
bio_list_init(&bio_list_on_stack);
spin_lock_irq(q->queue_lock);
blk_run_throtl_bios_weight(td, NULL, &bio_list_on_stack);
spin_unlock_irq(q->queue_lock);
if (!bio_list_empty(&bio_list_on_stack)) {
blk_start_plug(&plug);
/*lint -save -e820*/
while((bio = bio_list_pop(&bio_list_on_stack)))
/*lint -restore*/
generic_make_request(bio);
blk_finish_plug(&plug);
}
}
static void blk_throtl_rescue_timer_fn(unsigned long arg)
{
struct throtl_data *td = (struct throtl_data *)arg;
struct request_queue *q = td->queue;
unsigned long flags;
/*lint -save -e50 -e438 -e529 -e550 -e1058*/
if (ACCESS_ONCE(td->nr_queued[0])) {
smp_mb__after_atomic();
if (!atomic_read(&td->inflights))
queue_work(kthrotld_workqueue, &td->dispatch_work);
mod_timer(&td->rescue_timer,
jiffies + msecs_to_jiffies(THROTL_IDLE_INTERVAL));
return;
}
spin_lock_irqsave(q->queue_lock, flags);
if (ACCESS_ONCE(td->nr_queued[0])) {
smp_mb__after_atomic();
if (!atomic_read(&td->inflights))
queue_work(kthrotld_workqueue, &td->dispatch_work);
mod_timer(&td->rescue_timer,
jiffies + msecs_to_jiffies(THROTL_IDLE_INTERVAL));
}
spin_unlock_irqrestore(q->queue_lock, flags);
/*lint -restore*/
}
static void blk_throtl_idle_timer_fn(unsigned long arg)
{
struct throtl_data *td = (struct throtl_data *)arg;
struct request_queue *q = td->queue;
unsigned long flags;
/*lint -save -e550*/
spin_lock_irqsave(q->queue_lock, flags);
/*lint -restore*/
if (td->nr_queued[0])
queue_work(kthrotld_workqueue, &td->dispatch_work);
spin_unlock_irqrestore(q->queue_lock, flags);
}
static void tg_drain_bios_weight(struct throtl_grp *tg, struct bio_list *bios)
{
/*lint -save -e730 -e747*/
WARN_ON(list_empty(&tg->node));
list_del_init(&tg->node);
while (!bio_list_empty(&tg->bios))
tg_dispatch_one_bio_weight(tg, bios, false);
/*lint -restore*/
tg->io_cost.bytes_disp[0] = 0;
tg->io_cost.io_disp[0] = 0;
blkg_put(tg_to_blkg(tg));
}
static void blk_throtl_drain_weight(struct request_queue *q)
{
struct throtl_data *td = q->td;
struct throtl_grp *curr, *next;
struct bio *bio;
struct bio_list bios;
bio_list_init(&bios);
/*lint -save -e64 -e530 -e826*/
list_for_each_entry_safe(curr, next, &td->active, node)
tg_drain_bios_weight(curr, &bios);
list_for_each_entry_safe(curr, next, &td->expired, node)
tg_drain_bios_weight(curr, &bios);
/*lint -restore*/
spin_unlock_irq(q->queue_lock);
/*lint -save -e820*/
while((bio = bio_list_pop(&bios)))
generic_make_request(bio);
/*lint -restore*/
spin_lock_irq(q->queue_lock);
}
static void tg_drain_node_weight(struct throtl_grp *tg)
{
WARN_ON(list_empty(&tg->node));
list_del_init(&tg->node);
tg->io_cost.bytes_disp[0] = 0;
tg->io_cost.io_disp[0] = 0;
tg->expired = false;
wake_up_all(&tg->wait);
blkg_put(tg_to_blkg(tg));
}
static void __blk_throtl_drain_weight(struct request_queue *q)
{
struct throtl_data *td = q->td;
struct throtl_grp *curr, *next;
list_for_each_entry_safe(curr, next, &td->active, node)
tg_drain_node_weight(curr);
list_for_each_entry_safe(curr, next, &td->expired, node)
tg_drain_node_weight(curr);
}
static bool atomic_inc_below(atomic_t *v, int below)
{
/*lint -save -e438 -e529*/
int cur = atomic_read(v);
/*lint -restore*/
for (;;) {
int old;
if (cur >= below)
return false;
/*lint -save -e571*/
old = atomic_cmpxchg(v, cur, cur + 1);
if (old == cur)
break;
/*lint -restore*/
cur = old;
}
return true;
}
static bool io_event_ok(struct throtl_io_limit *io_limit, int limit)
{
if (limit == 0)
limit = INT_MAX;
return atomic_inc_below(&io_limit->inflights, limit);
}
void blk_throtl_update_limit(struct blk_throtl_io_limit *io_limit,
int limit)
{
io_limit->max_inflights = limit;
wake_up_all(&io_limit->io_limits[0].waitq);
wake_up_all(&io_limit->io_limits[1].waitq);
}
static void __throtl_limit_io_done(struct throtl_io_limit *io_limit,
int limit)
{
int inflights;
inflights = atomic_dec_return(&io_limit->inflights);
if (inflights >= limit) {
if (!limit && waitqueue_active(&io_limit->waitq))
wake_up_all(&io_limit->waitq);
return;
}
if (waitqueue_active(&io_limit->waitq)) {
int diff = max(limit - inflights, 1);
wake_up_nr(&io_limit->waitq, diff);
}
}
static void blk_throtl_limit_io_done(struct bio *bio)
{
struct blk_throtl_io_limit *io_limit;
int limit;
int index = (int)bio_data_dir(bio);
io_limit = (struct blk_throtl_io_limit *)bio->bi_throtl_private1;
/*lint -save -e438 -e529*/
limit = ACCESS_ONCE(io_limit->max_inflights);
/*lint -restore*/
__throtl_limit_io_done(&io_limit->io_limits[index], limit);
bio->bi_throtl_private1 = NULL;
bio->bi_throtl_end_io1 = NULL;
blk_throtl_io_limit_put(io_limit);
}
/* We must hold rcu lock. */
static int blk_throtl_io_limit_wait(struct blkcg *blkcg,
struct blk_throtl_io_limit *task_limit,
struct throtl_grp *tg,
struct bio *bio)
{
/*lint -save -e438 -e446 -e529 -e727 -e730*/
struct blkcg *new;
DEFINE_WAIT(wait);
int limit;
int index = (int)bio_data_dir(bio);
struct throtl_io_limit *io_limit = &task_limit->io_limits[index];
int ret;
WARN_ON_ONCE(!rcu_read_lock_held());
limit = ACCESS_ONCE(task_limit->max_inflights);
/*lint -restore*/
if (io_event_ok(io_limit, limit))
return 0;
ret = 0;
for (;;) {
prepare_to_wait(&io_limit->waitq, &wait, TASK_UNINTERRUPTIBLE);
/*lint -save -e438 -e529*/
limit = ACCESS_ONCE(task_limit->max_inflights);
/*lint -restore*/
new = bio_blkcg(bio);
if (new != blkcg) {
ret = 1;
break;
}
if (io_event_ok(io_limit, limit))
break;
if (!tg->td->max_inflights) {
ret = 2;
break;
}
task_set_sleep_on_throtl(current);
rcu_read_unlock();
#ifdef CONFIG_HUAWEI_IO_TRACING
trace_block_throttle_limit_start(bio, tg->td->max_inflights,
tg->inflights[bio_data_dir(bio)]);
#endif
io_schedule();
#ifdef CONFIG_HUAWEI_IO_TRACING
trace_block_throttle_limit_end(bio);
#endif
rcu_read_lock();
task_clear_sleep_on_throtl(current);
}
finish_wait(&io_limit->waitq, &wait);
return ret;
}
static int tg_print_cgroup_type(struct seq_file *sf, void *v)
{
struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
seq_printf(sf, "type: %d\n", blkcg->type);
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_set_cgroup_type(struct cgroup_subsys_state *css,
struct cftype *cft, u64 val)
{
struct blkcg *blkcg = css_to_blkcg(css);
unsigned int type = (unsigned int)val;
if (type >= BLK_THROTL_TYPE_NR)
return -EINVAL;
blkcg->type = type;
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static u64 tg_prfill_conf_u64(struct seq_file *sf, struct blkg_policy_data *pd,
int off)
{
struct throtl_grp *tg = pd_to_tg(pd);
/*lint -save -e124*/
u64 v = *(u64 *)((void *)tg + off);
/*lint -restore*/
/*lint -save -e650 -e737 -e747*/
if (v == -1)
return 0;
return __blkg_prfill_u64(sf, pd, v);
/*lint -restore*/
}
static u64 tg_prfill_conf_uint(struct seq_file *sf, struct blkg_policy_data *pd,
int off)
{
struct throtl_grp *tg = pd_to_tg(pd);
/*lint -save -e124*/
unsigned int v = *(unsigned int *)((void *)tg + off);
/*lint -restore*/
/*lint -save -e650 -e737 -e747*/
if (v == -1)
return 0;
return __blkg_prfill_u64(sf, pd, v);
/*lint -restore*/
}
static int tg_print_conf_u64(struct seq_file *sf, void *v)
{
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), tg_prfill_conf_u64,
&blkcg_policy_throtl, seq_cft(sf)->private, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_conf_uint(struct seq_file *sf, void *v)
{
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), tg_prfill_conf_uint,
&blkcg_policy_throtl, seq_cft(sf)->private, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static void tg_conf_updated(struct throtl_grp *tg)
{
struct throtl_service_queue *sq = &tg->service_queue;
struct cgroup_subsys_state *pos_css;
struct blkcg_gq *blkg;
throtl_log(&tg->service_queue,
"limit change rbps=%llu wbps=%llu riops=%u wiops=%u",
tg->io_cost.bps[READ], tg->io_cost.bps[WRITE],
tg->io_cost.iops[READ], tg->io_cost.iops[WRITE]);
/*
* Update has_rules[] flags for the updated tg's subtree. A tg is
* considered to have rules if either the tg itself or any of its
* ancestors has rules. This identifies groups without any
* restrictions in the whole hierarchy and allows them to bypass
* blk-throttle.
*/
/*lint -save -e666*/
blkg_for_each_descendant_pre(blkg, pos_css, tg_to_blkg(tg))
tg_update_has_rules(blkg_to_tg(blkg));
/*lint -restore*/
/*
* We're already holding queue_lock and know @tg is valid. Let's
* apply the new config directly.
*
* Restart the slices for both READ and WRITES. It might happen
* that a group's limit are dropped suddenly and we don't want to
* account recently dispatched IO with new low rate.
*/
throtl_start_new_slice(tg, 0);
throtl_start_new_slice(tg, 1);
if (tg->flags & THROTL_TG_PENDING) {
tg_update_disptime(tg);
throtl_schedule_next_dispatch(sq->parent_sq, true);
}
}
static ssize_t tg_set_conf(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off, bool is_u64)
{
struct blkcg *blkcg = css_to_blkcg(of_css(of));
struct blkg_conf_ctx ctx;
struct throtl_grp *tg;
int ret;
u64 v;
/*lint -save -e570*/
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
if (ret)
return ret;
ret = -EINVAL;
if (sscanf(ctx.body, "%llu", &v) != 1)
goto out_finish;
if (!v)
v = -1;
tg = blkg_to_tg(ctx.blkg);
if (td_weight_based(tg->td)) {
ret = -EBUSY;
goto out_finish;
}
/*lint -restore*/
if (is_u64)
*(u64 *)((void *)tg + of_cft(of)->private) = v;
else
*(unsigned int *)((void *)tg + of_cft(of)->private) = v;
tg->td->mode = MODE_THROTTLE;
tg_conf_updated(tg);
ret = 0;
out_finish:
blkg_conf_finish(&ctx);
return ret ?: nbytes;
/*lint -save -e715*/
}
/*lint -restore*/
static ssize_t tg_set_conf_u64(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
/*lint -save -e747*/
return tg_set_conf(of, buf, nbytes, off, true);
/*lint -restore*/
}
static ssize_t tg_set_conf_uint(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
/*lint -save -e747*/
return tg_set_conf(of, buf, nbytes, off, false);
/*lint -restore*/
}
static u64 tg_prfill_mode_device(struct seq_file *sf,
struct blkg_policy_data *pd, int off)
{
struct throtl_grp *tg = pd_to_tg(pd);
const char *dname = blkg_dev_name(pd->blkg);
if (!dname)
return 0;
if (tg->td->mode == MODE_NONE)
return 0;
seq_printf(sf, "%s %s\n", dname, run_mode_name[tg->td->mode]);
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int throtl_print_mode_device(struct seq_file *sf, void *v)
{
int i;
seq_printf(sf, "available ");
for (i = 0; i < MAX_MODE; i++)
seq_printf(sf, "%s(%d) ", run_mode_name[i], i);
seq_printf(sf, "\n");
seq_printf(sf, "default %s\n", run_mode_name[MODE_NONE]);
mutex_lock(&throtl_mode_lock);
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
tg_prfill_mode_device, &blkcg_policy_throtl, 0, false);
/*lint -restore*/
mutex_unlock(&throtl_mode_lock);
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static void tg_update_rules(struct throtl_data *td)
{
struct cgroup_subsys_state *pos_css;
struct blkcg_gq *blkg, *top_blkg;
top_blkg = td->queue->root_blkg;
/*lint -save -e747 -e820*/
blkg_for_each_descendant_pre(blkg, pos_css, top_blkg)
tg_update_has_rules(blkg_to_tg(blkg));
/*lint -restore*/
}
static void tg_change_mode(struct throtl_data *td, int mode)
{
struct request_queue *q = td->queue;
blk_queue_bypass_start(q);
/*
* Maybe it is unnecessary, but we need make sure all bios in the
* block throttle layer are drained.
*/
spin_lock_irq(q->queue_lock);
blkcg_drain_queue(q);
spin_unlock_irq(q->queue_lock);
/*lint -save -e40 -e64 -e438 -e529 -e666 -e730 -e774 -e845*/
WARN_ON(td->nr_queued[0]);
WARN_ON(td->nr_queued[1]);
wait_event(td->waitq, !atomic_read(&td->inflights));
del_timer_sync(&td->rescue_timer);
del_timer_sync(&td->idle_timer);
flush_work(&td->dispatch_work);
spin_lock_irq(q->queue_lock);
td->mode = mode;
/*lint -restore*/
if (td_weight_based(td))
INIT_WORK(&td->dispatch_work, blk_throtl_kick_queue_fn);
else
INIT_WORK(&td->dispatch_work, blk_throtl_dispatch_work_fn);
rcu_read_lock();
tg_update_share(td, NULL);
tg_update_rules(td);
rcu_read_unlock();
spin_unlock_irq(q->queue_lock);
blk_queue_bypass_end(q);
}
static ssize_t tg_set_mode_device(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct blkcg *blkcg = css_to_blkcg(of_css(of));
struct blkg_conf_ctx ctx;
struct throtl_grp *tg;
struct throtl_data *td;
struct gendisk *disk;
int ret;
int mode;
bool need_change = false;
mutex_lock(&throtl_mode_lock);
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
if (ret)
goto out;
if (sscanf(ctx.body, "%d", &mode) != 1)
goto out_finish;
ret = -EINVAL;
if (mode < 0 || mode >= MAX_MODE)
goto out_finish;
tg = blkg_to_tg(ctx.blkg);
if (tg->td->mode == mode) {
ret = 0;
goto out_finish;
}
disk = ctx.disk;
if (!get_disk(disk)) {
ret = -EBUSY;
goto out_finish;
}
td = tg->td;
need_change = true;
ret = 0;
out_finish:
blkg_conf_finish(&ctx);
/*lint -save -e644 -e713 -e737*/
if (need_change) {
tg_change_mode(td, mode);
put_disk(disk);
}
out:
mutex_unlock(&throtl_mode_lock);
return ret ?: nbytes;
/*lint -restore*/
/*lint -save -e715*/
}
/*lint -restore*/
static u64 tg_prfill_weight_device(struct seq_file *sf,
struct blkg_policy_data *pd, int off)
{
struct blkcg_gq *blkg = pd_to_blkg(pd);
const char *dname = blkg_dev_name(pd->blkg);
if (!dname)
return 0;
if (!blkg)
return 0;
if (blkg->weight == blkg->blkcg->weight)
return 0;
seq_printf(sf, "%s: weight %u\n", dname, blkg->weight);
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_weight(struct seq_file *sf, void *v)
{
struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
seq_printf(sf, "default: %u\n", blkcg->weight);
/*lint -save -e747*/
blkcg_print_blkgs(sf, blkcg, tg_prfill_weight_device,
&blkcg_policy_throtl, 0, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_weight_device(struct seq_file *sf, void *v)
{
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
tg_prfill_weight_device,
&blkcg_policy_throtl, 0, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static ssize_t tg_set_weight_device(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct blkcg *blkcg = css_to_blkcg(of_css(of));
struct throtl_grp *tg;
struct throtl_service_queue *parent_sq;
struct blkg_conf_ctx ctx;
unsigned int val;
int ret;
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
if (ret)
return ret;
ret = -EINVAL;
if (sscanf(ctx.body, "%d", &val) != 1)
goto out_finish;
if (val < BLKIO_WEIGHT_MIN || val > BLKIO_WEIGHT_MAX)
goto out_finish;
tg = blkg_to_tg(ctx.blkg);
parent_sq = tg->service_queue.parent_sq;
if (parent_sq) {
parent_sq->children_weight -= ctx.blkg->weight;
parent_sq->children_weight += val;
}
ctx.blkg->weight = val;
tg_update_share(tg->td, tg);
ret = 0;
out_finish:
blkg_conf_finish(&ctx);
/*lint -save -e713 -e737*/
return ret ?: nbytes;
/*lint -restore*/
/*lint -save -e715*/
}
/*lint -restore*/
static ssize_t tg_set_weight_offon(struct kernfs_open_file *of,
char *buf, size_t nbytes,
loff_t off)
{
int ret;
char offon[8];
if (sscanf(buf, "%7s", offon) != 1)
return -EINVAL;
ret = 0;
if (strnlen(offon, (size_t)7) == 5 &&
!strncmp(offon, "on-fs", (size_t)6))
blk_throtl_weight_offon = BLK_THROTL_WEIGHT_ON_FS;
else if (strnlen(offon, (size_t)7) == 6 &&
!strncmp(offon, "on-blk", (size_t)7))
blk_throtl_weight_offon = BLK_THROTL_WEIGHT_ON_BLK;
else if (strnlen(offon, (size_t)4) == 3 &&
!strncmp(offon, "off", (size_t)4))
blk_throtl_weight_offon = BLK_THROTL_WEIGHT_OFF;
else
ret = -EINVAL;
/*lint -save -e713 -e737*/
return ret ?: nbytes;
/*lint -restore*/
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_weight_offon(struct seq_file *sf, void *v)
{
seq_printf(sf, "Weight Based Throttle: %s\n",
(blk_throtl_weight_offon == BLK_THROTL_WEIGHT_ON_FS) ?
"on-fs" :
((blk_throtl_weight_offon == BLK_THROTL_WEIGHT_ON_BLK) ?
"on-blk" : "off"));
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static u64 tg_prfill_max_inflights_device(struct seq_file *sf,
struct blkg_policy_data *pd, int off)
{
struct throtl_grp *tg = pd_to_tg(pd);
const char *dname = blkg_dev_name(pd->blkg);
if (!dname)
return 0;
/*lint -save -e438 -e529*/
seq_printf(sf, "%s: queued %u, read inflight %d, write inflight %d, devices inflights %d delay count %d, max delay %d\n\n",
dname, tg->service_queue.nr_queued[0],
atomic_read(&tg->inflights[0]),
atomic_read(&tg->inflights[1]),
atomic_read(&tg->td->inflights),
atomic_read(&tg->delay_cnt),
atomic_read(&tg->max_delay));
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_max_inflights(struct seq_file *sf, void *v)
{
struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
seq_printf(sf, "max limit: %d\n", blkcg->max_inflights);
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
tg_prfill_max_inflights_device,
&blkcg_policy_throtl, 0, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_set_max_inflights(struct cgroup_subsys_state *css,
struct cftype *cft, s64 val)
{
struct blkcg *blkcg = css_to_blkcg(css);
struct task_struct *task;
struct css_task_iter it;
int max_inflights = (int)val;
if (max_inflights < MIN_INFLIGHT || max_inflights > MAX_INFLIGHT)
return -EINVAL;
spin_lock_irq(&blkcg->lock);
if (blkcg->max_inflights == max_inflights)
goto out_unlock;
blkcg->max_inflights = max_inflights;
rcu_read_lock();
css_task_iter_start(css, &it);
/*lint -save -e820*/
while ((task = css_task_iter_next(&it))) {
/*lint -restore*/
if (!task->mm)
continue;
if (!task->mm->io_limit)
continue;
blk_throtl_update_limit(task->mm->io_limit, max_inflights);
}
css_task_iter_end(&it);
rcu_read_unlock();
out_unlock:
spin_unlock_irq(&blkcg->lock);
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static u64 tg_prfill_enable_max_inflights_device(struct seq_file *sf,
struct blkg_policy_data *pd, int off)
{
struct throtl_grp *tg = pd_to_tg(pd);
const char *dname = blkg_dev_name(pd->blkg);
if (!dname)
return 0;
seq_printf(sf, "%s: limit %s\n",
dname, (tg->td->max_inflights ? "open" : "close"));
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_enable_max_inflight_device(struct seq_file *sf, void *v)
{
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
tg_prfill_enable_max_inflights_device,
&blkcg_policy_throtl, 0, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static ssize_t tg_set_enable_max_inflights_device(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct cgroup_subsys_state *css = of_css(of);
struct blkcg *blkcg = css_to_blkcg(css);
struct blkcg_gq *blkg;
struct blkg_conf_ctx ctx;
struct throtl_grp *tg;
struct task_struct *task;
struct css_task_iter it;
unsigned long flags;
int max_inflight;
int ret;
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
if (ret)
return ret;
ret = -EINVAL;
if (sscanf(ctx.body, "%d", &max_inflight) != 1)
goto out_finish;
if (max_inflight != 0 && max_inflight != 1)
goto out_finish;
ret = 0;
tg = blkg_to_tg(ctx.blkg);
if (tg->td->max_inflights == max_inflight)
goto out_finish;
/*lint -save -e550*/
spin_lock_irqsave(&blkcg->lock, flags);
/*lint -restore*/
if (!max_inflight)
goto skip_check;
/* we just open only one device for limit */
/*lint -save -e62 -e64 -e826*/
hlist_for_each_entry(blkg, &blkcg->blkg_list, blkcg_node) {
/*lint -restore*/
struct throtl_grp *curr_tg = blkg_to_tg(blkg);
if (!curr_tg || curr_tg == tg)
continue;
if (curr_tg->td->max_inflights) {
ret = -EBUSY;
goto out_unlock;
}
}
skip_check:
tg->td->max_inflights = max_inflight;
smp_mb();
if (max_inflight)
goto out_unlock;
rcu_read_lock();
css_task_iter_start(css, &it);
/*lint -save -e820*/
while ((task = css_task_iter_next(&it))) {
/*lint -restore*/
if (task_sleep_on_throtl(task))
wake_up_process(task);
}
css_task_iter_end(&it);
rcu_read_unlock();
out_unlock:
spin_unlock_irqrestore(&blkcg->lock, flags);
out_finish:
blkg_conf_finish(&ctx);
/*lint -save -e713 -e737*/
return ret ?: nbytes;
/*lint -restore*/
/*lint -save -e715*/
}
/*lint -restore*/
static u64 tg_prfill_quantum_device(struct seq_file *sf,
struct blkg_policy_data *pd, int off)
{
struct blkcg_gq *blkg = pd_to_blkg(pd);
const char *dname = blkg_dev_name(pd->blkg);
if (!dname)
return 0;
if (!blkg)
return 0;
if ((blkg_to_tg(blkg)->td->quantum) == QUANTUM_INQUEUE_DEF)
return 0;
seq_printf(sf, "%s: quantum %u\n", dname,
blkg_to_tg(blkg)->td->quantum);
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_quantum_device(struct seq_file *sf, void *v)
{
seq_printf(sf, "default: %d\n", QUANTUM_INQUEUE_DEF);
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
tg_prfill_quantum_device,
&blkcg_policy_throtl, 0, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static ssize_t tg_set_quantum_device(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct blkcg *blkcg = css_to_blkcg(of_css(of));
struct throtl_grp *tg;
struct blkg_conf_ctx ctx;
int val;
int ret;
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
if (ret)
return ret;
ret = -EINVAL;
if (sscanf(ctx.body, "%d", &val) != 1)
goto out_finish;
if (val < QUANTUM_INQUEUE_MIN || val > QUANTUM_INQUEUE_MAX)
goto out_finish;
tg = blkg_to_tg(ctx.blkg);
tg->td->quantum = val;
ret = 0;
out_finish:
blkg_conf_finish(&ctx);
/*lint -save -e713 -e737*/
return ret ?: nbytes;
/*lint -restore*/
/*lint -save -e715*/
}
/*lint -restore*/
static u64 tg_prfill_bw_slice_device(struct seq_file *sf,
struct blkg_policy_data *pd, int off)
{
struct blkcg_gq *blkg = pd_to_blkg(pd);
const char *dname = blkg_dev_name(pd->blkg);
if (!dname)
return 0;
if (!blkg)
return 0;
if ((blkg_to_tg(blkg)->td->bw_slice) == BW_SLICE_DEF)
return 0;
seq_printf(sf, "%s: bw_slice %llu\n", dname,
blkg_to_tg(blkg)->td->bw_slice);
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_bw_slice_device(struct seq_file *sf, void *v)
{
seq_printf(sf, "default: %llu\n", BW_SLICE_DEF);
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
tg_prfill_bw_slice_device,
&blkcg_policy_throtl, 0, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static ssize_t tg_set_bw_slice_device(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct blkcg *blkcg = css_to_blkcg(of_css(of));
struct throtl_grp *tg;
struct blkg_conf_ctx ctx;
uint64_t val;
int ret;
/*lint -save -e561*/
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
if (ret)
return ret;
ret = -EINVAL;
if (sscanf(ctx.body, "%llu", &val) != 1)
goto out_finish;
if (val < BW_SLICE_MIN || val > BW_SLICE_MAX)
goto out_finish;
tg = blkg_to_tg(ctx.blkg);
tg->td->bw_slice = val;
tg_update_share(tg->td, NULL);
/*lint -restore*/
ret = 0;
out_finish:
blkg_conf_finish(&ctx);
/*lint -save -e713 -e737*/
return ret ?: nbytes;
/*lint -restore*/
/*lint -save -e715*/
}
/*lint -restore*/
static u64 tg_prfill_iops_slice_device(struct seq_file *sf,
struct blkg_policy_data *pd,
int off)
{
struct blkcg_gq *blkg = pd_to_blkg(pd);
const char *dname = blkg_dev_name(pd->blkg);
if (!dname)
return 0;
if (!blkg)
return 0;
if (blkg_to_tg(blkg)->td->iops_slice == IOPS_SLICE_DEF)
return 0;
seq_printf(sf, "%s: iops_slice %u\n", dname,
blkg_to_tg(blkg)->td->iops_slice);
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static int tg_print_iops_slice_device(struct seq_file *sf, void *v)
{
seq_printf(sf, "default: %d\n", IOPS_SLICE_DEF);
/*lint -save -e747*/
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
tg_prfill_iops_slice_device,
&blkcg_policy_throtl, 0, false);
/*lint -restore*/
return 0;
/*lint -save -e715*/
}
/*lint -restore*/
static ssize_t tg_set_iops_slice_device(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct blkcg *blkcg = css_to_blkcg(of_css(of));
struct throtl_grp *tg;
struct blkg_conf_ctx ctx;
unsigned int val;
int ret;
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
if (ret)
return ret;
ret = -EINVAL;
if (sscanf(ctx.body, "%d", &val) != 1)
goto out_finish;
if (val < IOPS_SLICE_MIN || val > IOPS_SLICE_MAX)
goto out_finish;
tg = blkg_to_tg(ctx.blkg);
tg->td->iops_slice = val;
tg_update_share(tg->td, NULL);
ret = 0;
out_finish:
blkg_conf_finish(&ctx);
/*lint -save -e713 -e737*/
return ret ?: nbytes;
/*lint -restore*/
/*lint -save -e715*/
}
/*lint -restore*/
/*lint -save -e785*/
static struct cftype throtl_legacy_files[] = {
{
.name = "throttle.mode_device",
.flags = CFTYPE_ONLY_ON_ROOT,
.seq_show = throtl_print_mode_device,
.write = tg_set_mode_device,
},
{
.name = "throttle.read_bps_device",
.private = offsetof(struct throtl_grp, io_cost.bps[READ]),
.seq_show = tg_print_conf_u64,
.write = tg_set_conf_u64,
},
{
.name = "throttle.write_bps_device",
.private = offsetof(struct throtl_grp, io_cost.bps[WRITE]),
.seq_show = tg_print_conf_u64,
.write = tg_set_conf_u64,
},
{
.name = "throttle.read_iops_device",
.private = offsetof(struct throtl_grp, io_cost.iops[READ]),
.seq_show = tg_print_conf_uint,
.write = tg_set_conf_uint,
},
{
.name = "throttle.write_iops_device",
.private = offsetof(struct throtl_grp, io_cost.iops[WRITE]),
.seq_show = tg_print_conf_uint,
.write = tg_set_conf_uint,
},
{
.name = "throttle.io_service_bytes",
.private = (unsigned long)&blkcg_policy_throtl,
.seq_show = blkg_print_stat_bytes,
},
{
.name = "throttle.io_serviced",
.private = (unsigned long)&blkcg_policy_throtl,
.seq_show = blkg_print_stat_ios,
},
{
.name = "throttle.weight_device",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = tg_print_weight_device,
.write = tg_set_weight_device,
},
{
.name = "throttle.weight",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = tg_print_weight,
},
{
.name = "throttle.weight_based_offon",
.flags = CFTYPE_ONLY_ON_ROOT,
.seq_show = tg_print_weight_offon,
.write = tg_set_weight_offon,
},
{
.name = "throttle.max_inflights",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = tg_print_max_inflights,
.write_s64 = tg_set_max_inflights,
},
{
.name = "throttle.enable_max_inflights_device",
.flags = CFTYPE_ONLY_ON_ROOT,
.seq_show = tg_print_enable_max_inflight_device,
.write = tg_set_enable_max_inflights_device,
},
{
.name = "throttle.quantum_device",
.flags = CFTYPE_ONLY_ON_ROOT,
.seq_show = tg_print_quantum_device,
.write = tg_set_quantum_device,
},
{
.name = "throttle.bw_slice_device",
.flags = CFTYPE_ONLY_ON_ROOT,
.seq_show = tg_print_bw_slice_device,
.write = tg_set_bw_slice_device,
},
{
.name = "throttle.iops_slice_device",
.flags = CFTYPE_ONLY_ON_ROOT,
.seq_show = tg_print_iops_slice_device,
.write = tg_set_iops_slice_device,
},
{
.name = "throttle.type",
.seq_show = tg_print_cgroup_type,
.write_u64 = tg_set_cgroup_type,
},
{ } /* terminate */
};
/*lint -restore*/
static u64 tg_prfill_max(struct seq_file *sf, struct blkg_policy_data *pd,
int off)
{
struct throtl_grp *tg = pd_to_tg(pd);
const char *dname = blkg_dev_name(pd->blkg);
char bufs[4][21] = { "max", "max", "max", "max" };
/*lint -save -e650*/
if (!dname)
return 0;
if (tg->io_cost.bps[READ] == -1 && tg->io_cost.bps[WRITE] == -1 &&
tg->io_cost.iops[READ] == -1 && tg->io_cost.iops[WRITE] == -1)
return 0;
if (tg->io_cost.bps[READ] != -1)
snprintf(bufs[0], sizeof(bufs[0]), "%llu", tg->io_cost.bps[READ]);
if (tg->io_cost.bps[WRITE] != -1)
snprintf(bufs[1], sizeof(bufs[1]), "%llu", tg->io_cost.bps[WRITE]);
if (tg->io_cost.iops[READ] != -1)
snprintf(bufs[2], sizeof(bufs[2]), "%u", tg->io_cost.iops[READ]);
if (tg->io_cost.iops[WRITE] != -1)
snprintf(bufs[3], sizeof(bufs[3]), "%u", tg->io_cost.iops[WRITE]);
/*lint -restore*/
seq_printf(sf, "%s rbps=%s wbps=%s riops=%s wiops=%s\n",
dname, bufs[0], bufs[1], bufs[2], bufs[3]);
return 0;
}
static int tg_print_max(struct seq_file *sf, void *v)
{
blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), tg_prfill_max,
&blkcg_policy_throtl, seq_cft(sf)->private, false);
return 0;
}
static ssize_t tg_set_max(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct blkcg *blkcg = css_to_blkcg(of_css(of));
struct blkg_conf_ctx ctx;
struct throtl_grp *tg;
u64 v[4];
int ret;
ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
if (ret)
return ret;
tg = blkg_to_tg(ctx.blkg);
v[0] = tg->io_cost.bps[READ];
v[1] = tg->io_cost.bps[WRITE];
v[2] = tg->io_cost.iops[READ];
v[3] = tg->io_cost.iops[WRITE];
/*lint -save -e570*/
while (true) {
char tok[27]; /* wiops=18446744073709551616 */
char *p;
u64 val = -1;
int len;
/*lint -restore*/
if (sscanf(ctx.body, "%26s%n", tok, &len) != 1)
break;
if (tok[0] == '\0')
break;
ctx.body += len;
/*lint -save -e421*/
ret = -EINVAL;
p = tok;
strsep(&p, "=");
if (!p || (sscanf(p, "%llu", &val) != 1 && strcmp(p, "max")))
goto out_finish;
ret = -ERANGE;
if (!val)
goto out_finish;
ret = -EINVAL;
if (!strcmp(tok, "rbps"))
v[0] = val;
else if (!strcmp(tok, "wbps"))
v[1] = val;
else if (!strcmp(tok, "riops"))
v[2] = min_t(u64, val, UINT_MAX);
else if (!strcmp(tok, "wiops"))
v[3] = min_t(u64, val, UINT_MAX);
else
goto out_finish;
}
/*lint -restore*/
tg->io_cost.bps[READ] = v[0];
tg->io_cost.bps[WRITE] = v[1];
tg->io_cost.iops[READ] = v[2];
tg->io_cost.iops[WRITE] = v[3];
tg_conf_updated(tg);
ret = 0;
out_finish:
blkg_conf_finish(&ctx);
return ret ?: nbytes;
}
static struct cftype throtl_files[] = {
{
.name = "max",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = tg_print_max,
.write = tg_set_max,
},
{ } /* terminate */
};
static void throtl_shutdown_wq(struct request_queue *q)
{
struct throtl_data *td = q->td;
del_timer_sync(&td->rescue_timer);
del_timer_sync(&td->idle_timer);
cancel_work_sync(&td->dispatch_work);
}
/*lint -save -e31 -e64*/
static struct blkcg_policy blkcg_policy_throtl = {
.dfl_cftypes = throtl_files,
.legacy_cftypes = throtl_legacy_files,
.pd_alloc_fn = throtl_pd_alloc,
.pd_init_fn = throtl_pd_init,
.pd_online_fn = throtl_pd_online,
.pd_offline_fn = throtl_pd_offline,
.pd_free_fn = throtl_pd_free,
};
/*lint -restore*/
bool blk_throtl_bio(struct request_queue *q, struct blkcg_gq *blkg,
struct bio *bio)
{
struct throtl_qnode *qn = NULL;
struct throtl_grp *tg = blkg_to_tg(blkg ?: q->root_blkg);
struct throtl_service_queue *sq;
bool rw = bio_data_dir(bio);
int index;
struct throtl_data *td = q->td;
struct blkcg *blkcg;
bool throttled = false;
WARN_ON_ONCE(!rcu_read_lock_held());
/* see throtl_charge_bio() */
if ((bio->bi_opf & (REQ_THROTTLED |REQ_META | REQ_PREFLUSH |
REQ_FUA | REQ_OP_DISCARD | REQ_CHAINED)))
goto out;
/* If we are on the way of memory reclaim, skip throttle */
if (current->flags & (PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD))
goto out;
if (task_in_pagefault(current))
goto out;
again:
if (!blkg)
goto out;
index = tg_data_index(tg, rw);
if (!tg->has_rules[index])
goto out;
/*lint -save -e613*/
blkcg = blkg->blkcg;
if (td_weight_based(td) &&
(blkcg == &blkcg_root || !blk_throtl_weight_offon))
goto out;
/*lint -restore*/
if (blkcg->type <= BLK_THROTL_FG)
bio->bi_opf |= REQ_FG;
if (td_weight_based(td) &&
(blk_throtl_weight_offon == BLK_THROTL_WEIGHT_ON_FS))
goto out;
/*
* Now we just support limit control for one layer.
*/
/*lint -save -e438 -e529 -e666*/
if (ACCESS_ONCE(tg->td->max_inflights) && !bio->bi_throtl_private1 &&
current->mm &&
ACCESS_ONCE(current->mm->io_limit->max_inflights)) {
/*lint -restore*/
struct blk_throtl_io_limit *io_limit = current->mm->io_limit;
int ret;
blkg_get(tg_to_blkg(tg));
ret = blk_throtl_io_limit_wait(blkcg, io_limit, tg, bio);
blkg_put(tg_to_blkg(tg));
if (ret == 1) {
blkcg = bio_blkcg(bio);
blkg = blkg_lookup(blkcg, q);
if (unlikely(!blkg)) {
spin_lock_irq(q->queue_lock);
blkg = blkg_lookup_create(blkcg, q);
if (IS_ERR(blkg))
blkg = q->root_blkg;
spin_unlock_irq(q->queue_lock);
}
tg = blkg_to_tg(blkg ?: q->root_blkg);
goto again;
} else if (ret == 2)
goto skip;
blk_throtl_io_limit_get(io_limit);
bio->bi_throtl_private1 = io_limit;
bio->bi_throtl_end_io1 = blk_throtl_limit_io_done;
}
skip:
spin_lock_irq(q->queue_lock);
/*lint -save -e730*/
if (unlikely(blk_queue_bypass(q)))
/*lint -restore*/
goto out_unlock;
if (!(tg->flags & THROTL_TG_ONLINE))
goto out_unlock;
if (unlikely(blk_throtl_weight_offon != BLK_THROTL_WEIGHT_ON_BLK))
goto out_unlock;
if (td_weight_based(td)) {
struct bio_list bios;
/*
* Now we just support weight based throttle for
* one layer.
*/
if (bio->bi_throtl_end_io2)
goto out_unlock;
bio_list_init(&bios);
throttled = true;
bio_associate_current(bio);
throtl_add_bio_weight(bio, tg);
blk_run_throtl_bios_weight(td, tg, &bios);
spin_unlock_irq(q->queue_lock);
rcu_read_unlock();
/*lint -save -e820*/
while((bio = bio_list_pop(&bios)))
/*lint -restore*/
generic_make_request(bio);
rcu_read_lock();
goto out;
}
sq = &tg->service_queue;
while (true) {
/* throtl is FIFO - if bios are already queued, should queue */
if (sq->nr_queued[index])
break;
/* if above limits, break to queue */
if (!tg_may_dispatch(tg, bio, NULL))
break;
/* within limits, let's charge and dispatch directly */
throtl_charge_bio(tg, bio);
/*
* We need to trim slice even when bios are not being queued
* otherwise it might happen that a bio is not queued for
* a long time and slice keeps on extending and trim is not
* called for a long time. Now if limits are reduced suddenly
* we take into account all the IO dispatched so far at new
* low rate and * newly queued IO gets a really long dispatch
* time.
*
* So keep on trimming slice even if bio is not queued.
*/
throtl_trim_slice(tg, rw);
/*
* @bio passed through this layer without being throttled.
* Climb up the ladder. If we''re already at the top, it
* can be executed directly.
*/
qn = &tg->qnode_on_parent[index];
sq = sq->parent_sq;
tg = sq_to_tg(sq);
if (!tg)
goto out_unlock;
}
/* out-of-limit, queue to @tg */
/*lint -save -e747*/
throtl_log(sq, "[%c] bio. bdisp=%llu sz=%u bps=%llu iodisp=%u iops=%u queued=%d/%d",
rw == READ ? 'R' : 'W',
tg->io_cost.bytes_disp[index], bio->bi_iter.bi_size,
tg->io_cost.bps[index], tg->io_cost.io_disp[index],
tg->io_cost.iops[index],
sq->nr_queued[READ], sq->nr_queued[WRITE]);
/*lint -restore*/
bio_associate_current(bio);
tg->td->nr_queued[index]++;
throtl_add_bio_tg(bio, qn, tg);
throttled = true;
/*
* Update @tg's dispatch time and force schedule dispatch if @tg
* was empty before @bio. The forced scheduling isn't likely to
* cause undue delay as @bio is likely to be dispatched directly if
* its @tg's disptime is not in the future.
*/
if (tg->flags & THROTL_TG_WAS_EMPTY) {
tg_update_disptime(tg);
/*lint -save -e747*/
throtl_schedule_next_dispatch(tg->service_queue.parent_sq, true);
/*lint -restore*/
}
out_unlock:
spin_unlock_irq(q->queue_lock);
out:
/*
* As multiple blk-throtls may stack in the same issue path, we
* don't want bios to leave with the flag set. Clear the flag if
* being issued.
*/
if (!throttled)
/*lint -save -e613*/
bio->bi_opf &= ~REQ_THROTTLED;
/*lint -restore*/
return throttled;
}
static inline void blk_throtl_update_quota(struct throtl_grp *tg,
unsigned int size)
{
tg->intime = jiffies;
tg->io_cost.bytes_disp[0] += size;
tg->io_cost.io_disp[0]++;
}
bool blk_throtl_get_quota(struct block_device *bdev, unsigned int size,
unsigned long jiffies_time_out, bool wait)
{
struct request_queue *q;
struct throtl_data *td;
struct throtl_grp *tg;
struct blkcg *blkcg = NULL;
struct blkcg *new_blkcg;
unsigned long start_jiffies;
unsigned long wait_jiffies = jiffies_time_out;
int ret;
if (current->flags & (PF_MEMALLOC | PF_SWAPWRITE |
PF_KSWAPD | PF_MUTEX_GC))
return true;
/*
* bdev has been got at mount(), and will be freed at kill_sb() after
* all files closed and all dirty pages flushed. So it's safe to
* using bdev without lock here.
*/
if (unlikely(!bdev) || unlikely(!bdev->bd_disk))
return true;
q = bdev_get_queue(bdev);
if (unlikely(!q))
return true;
td = q->td;
start_jiffies = jiffies;
get_new_group:
/*
* A throtl_grp pointer retrieved under rcu can be used to access
* basic fields.
*/
rcu_read_lock();
if (!td_weight_based(td))
goto out_unlock_rcu;
if (blk_throtl_weight_offon != BLK_THROTL_WEIGHT_ON_FS)
goto out_unlock_rcu;
new_blkcg = task_blkcg(current);
if (new_blkcg == &blkcg_root)
goto out_unlock_rcu;
if (!blkcg || blkcg != new_blkcg) {
struct blkcg_gq *blkg;
blkcg = new_blkcg;
blkg = blkg_lookup(blkcg, q);
if (unlikely(!blkg)) {
spin_lock_irq(q->queue_lock);
blkg = blkg_lookup_create(blkcg, q);
if (IS_ERR(blkg))
blkg = q->root_blkg;
spin_unlock_irq(q->queue_lock);
}
tg = blkg_to_tg(blkg);
}
spin_lock_irq(q->queue_lock);
if (unlikely(blk_queue_bypass(q)))
goto out_unlock;
if (unlikely(!(tg->flags & THROTL_TG_ONLINE)))
goto out_unlock;
if (unlikely(!td_weight_based(td)))
goto out_unlock;
if (unlikely(blk_throtl_weight_offon != BLK_THROTL_WEIGHT_ON_FS))
goto out_unlock;
retry:
if (tg_may_dispatch_weight(tg)) {
blk_throtl_update_quota(tg, size);
if (!list_empty(&tg->node))
goto out_unlock;
blkg_get(tg_to_blkg(tg));
list_add_tail(&tg->node, &td->active);
goto out_unlock;
} else if (!wait) {
spin_unlock_irq(q->queue_lock);
rcu_read_unlock();
return false;
}
if (!tg->expired) {
tg->expired = true;
list_move_tail(&tg->node, &td->expired);
if (list_empty(&td->active))
goto start_new_slice;
}
/*
* If jiffies_time_out expires, just return to continue
* to read/write.
*/
wait_jiffies = jiffies_time_out - (jiffies - start_jiffies);
if (jiffies - start_jiffies >= jiffies_time_out)
goto out_unlock;
if (should_start_new_slice(td))
goto start_new_slice;
blkg_get(tg_to_blkg(tg));
spin_unlock_irq(q->queue_lock);
rcu_read_unlock();
/*lint -save -e665 -e666*/
/* Waiting for starting new slice, max waiting time is 20ms */
ret = wait_event_interruptible_timeout(tg->wait, !tg->expired,
wait_jiffies >
msecs_to_jiffies(THROTL_IDLE_INTERVAL / 2) ?
msecs_to_jiffies(THROTL_IDLE_INTERVAL / 2) :
wait_jiffies);
/*lint -restore*/
blkg_put(tg_to_blkg(tg));
/* If this task is interrupted, just return */
if (signal_pending_state(TASK_INTERRUPTIBLE, current))
return true;
/* this task may switch to other group, get tg again */
goto get_new_group;
start_new_slice:
/* Start new slice and wake up waiting tgs */
throtl_start_new_slice_weight(td);
goto retry;
out_unlock:
spin_unlock_irq(q->queue_lock);
out_unlock_rcu:
rcu_read_unlock();
return true;
}
/*
* Dispatch all bios from all children tg's queued on @parent_sq. On
* return, @parent_sq is guaranteed to not have any active children tg's
* and all bios from previously active tg's are on @parent_sq->bio_lists[].
*/
static void tg_drain_bios(struct throtl_service_queue *parent_sq)
{
struct throtl_grp *tg;
/*lint -save -e747 -e820*/
while ((tg = throtl_rb_first(parent_sq))) {
struct throtl_service_queue *sq = &tg->service_queue;
struct bio *bio;
throtl_dequeue_tg(tg);
while ((bio = throtl_peek_queued(&sq->queued[READ])))
tg_dispatch_one_bio(tg, bio_data_dir(bio));
while ((bio = throtl_peek_queued(&sq->queued[WRITE])))
tg_dispatch_one_bio(tg, bio_data_dir(bio));
}
/*lint -restore*/
}
/**
* blk_throtl_drain - drain throttled bios
* @q: request_queue to drain throttled bios for
*
* Dispatch all currently throttled bios on @q through ->make_request_fn().
*/
void blk_throtl_drain(struct request_queue *q)
__releases(q->queue_lock) __acquires(q->queue_lock)
{
struct throtl_data *td = q->td;
struct blkcg_gq *blkg;
struct cgroup_subsys_state *pos_css;
struct bio *bio;
int i;
if (td_weight_based(td)) {
if (blk_throtl_weight_offon == BLK_THROTL_WEIGHT_ON_FS)
__blk_throtl_drain_weight(q);
else
blk_throtl_drain_weight(q);
return;
}
queue_lockdep_assert_held(q);
rcu_read_lock();
/*
* Drain each tg while doing post-order walk on the blkg tree, so
* that all bios are propagated to td->service_queue. It'd be
* better to walk service_queue tree directly but blkg walk is
* easier.
*/
/*lint -save -e747 -e820*/
blkg_for_each_descendant_post(blkg, pos_css, td->queue->root_blkg)
tg_drain_bios(&blkg_to_tg(blkg)->service_queue);
/*lint -restore*/
/* finally, transfer bios from top-level tg's into the td */
tg_drain_bios(&td->service_queue);
rcu_read_unlock();
spin_unlock_irq(q->queue_lock);
/* all bios now should be in td->service_queue, issue them */
/*lint -save -e574 -e737 -e820*/
for (i = READ; i <= WRITE; i++)
while ((bio = throtl_pop_queued(&td->service_queue.queued[i],
NULL)))
generic_make_request(bio);
/*lint -restore*/
spin_lock_irq(q->queue_lock);
}
int blk_throtl_init(struct request_queue *q)
{
struct throtl_data *td;
int ret;
td = kzalloc_node(sizeof(*td), GFP_KERNEL, q->node);
if (!td)
return -ENOMEM;
INIT_LIST_HEAD(&td->active);
INIT_LIST_HEAD(&td->expired);
/*lint -save -e1058 */
atomic_set(&td->inflights, 0);
init_waitqueue_head(&td->waitq);
setup_timer(&td->rescue_timer, blk_throtl_rescue_timer_fn,
(unsigned long)td);
setup_timer(&td->idle_timer, blk_throtl_idle_timer_fn,
(unsigned long)td);
/*lint -restore*/
INIT_WORK(&td->dispatch_work, blk_throtl_dispatch_work_fn);
throtl_service_queue_init(&td->service_queue);
td->service_queue.share = MAX_SHARE;
td->mode = MODE_NONE;
td->quantum = QUANTUM_INQUEUE_DEF;
td->bw_slice = BW_SLICE_DEF;
td->iops_slice = IOPS_SLICE_DEF;
q->td = td;
td->queue = q;
/* activate policy */
ret = blkcg_activate_policy(q, &blkcg_policy_throtl);
if (ret)
kfree(td);
return ret;
}
void blk_throtl_exit(struct request_queue *q)
{
/*lint -save -e730*/
BUG_ON(!q->td);
/*lint -restore*/
throtl_shutdown_wq(q);
blkcg_deactivate_policy(q, &blkcg_policy_throtl);
kfree(q->td);
}
/*lint -e528 -esym(528,*)*/
static int __init throtl_init(void)
{
kthrotld_workqueue = alloc_workqueue("kthrotld", WQ_MEM_RECLAIM, 0);
if (!kthrotld_workqueue)
panic("Failed to create kthrotld\n");
return blkcg_policy_register(&blkcg_policy_throtl);
}
/*lint -e528 +esym(528,*)*/
module_init(throtl_init);
Reference in a new issue View git blame Copy permalink