diff options
74 files changed, 3890 insertions, 1242 deletions
diff --git a/Documentation/cgroups/blkio-controller.txt b/Documentation/cgroups/blkio-controller.txt index cd556b914786..68b6a6a470b0 100644 --- a/Documentation/cgroups/blkio-controller.txt +++ b/Documentation/cgroups/blkio-controller.txt @@ -387,8 +387,81 @@ groups and put applications in that group which are not driving enough IO to keep disk busy. In that case set group_idle=0, and CFQ will not idle on individual groups and throughput should improve. -What works -========== -- Currently only sync IO queues are support. All the buffered writes are - still system wide and not per group. Hence we will not see service - differentiation between buffered writes between groups. +Writeback +========= + +Page cache is dirtied through buffered writes and shared mmaps and +written asynchronously to the backing filesystem by the writeback +mechanism. Writeback sits between the memory and IO domains and +regulates the proportion of dirty memory by balancing dirtying and +write IOs. + +On traditional cgroup hierarchies, relationships between different +controllers cannot be established making it impossible for writeback +to operate accounting for cgroup resource restrictions and all +writeback IOs are attributed to the root cgroup. + +If both the blkio and memory controllers are used on the v2 hierarchy +and the filesystem supports cgroup writeback, writeback operations +correctly follow the resource restrictions imposed by both memory and +blkio controllers. + +Writeback examines both system-wide and per-cgroup dirty memory status +and enforces the more restrictive of the two. Also, writeback control +parameters which are absolute values - vm.dirty_bytes and +vm.dirty_background_bytes - are distributed across cgroups according +to their current writeback bandwidth. + +There's a peculiarity stemming from the discrepancy in ownership +granularity between memory controller and writeback. While memory +controller tracks ownership per page, writeback operates on inode +basis. cgroup writeback bridges the gap by tracking ownership by +inode but migrating ownership if too many foreign pages, pages which +don't match the current inode ownership, have been encountered while +writing back the inode. + +This is a conscious design choice as writeback operations are +inherently tied to inodes making strictly following page ownership +complicated and inefficient. The only use case which suffers from +this compromise is multiple cgroups concurrently dirtying disjoint +regions of the same inode, which is an unlikely use case and decided +to be unsupported. Note that as memory controller assigns page +ownership on the first use and doesn't update it until the page is +released, even if cgroup writeback strictly follows page ownership, +multiple cgroups dirtying overlapping areas wouldn't work as expected. +In general, write-sharing an inode across multiple cgroups is not well +supported. + +Filesystem support for cgroup writeback +--------------------------------------- + +A filesystem can make writeback IOs cgroup-aware by updating +address_space_operations->writepage[s]() to annotate bio's using the +following two functions. + +* wbc_init_bio(@wbc, @bio) + + Should be called for each bio carrying writeback data and associates + the bio with the inode's owner cgroup. Can be called anytime + between bio allocation and submission. + +* wbc_account_io(@wbc, @page, @bytes) + + Should be called for each data segment being written out. While + this function doesn't care exactly when it's called during the + writeback session, it's the easiest and most natural to call it as + data segments are added to a bio. + +With writeback bio's annotated, cgroup support can be enabled per +super_block by setting MS_CGROUPWB in ->s_flags. This allows for +selective disabling of cgroup writeback support which is helpful when +certain filesystem features, e.g. journaled data mode, are +incompatible. + +wbc_init_bio() binds the specified bio to its cgroup. Depending on +the configuration, the bio may be executed at a lower priority and if +the writeback session is holding shared resources, e.g. a journal +entry, may lead to priority inversion. There is no one easy solution +for the problem. Filesystems can try to work around specific problem +cases by skipping wbc_init_bio() or using bio_associate_blkcg() +directly. diff --git a/Documentation/cgroups/memory.txt b/Documentation/cgroups/memory.txt index f456b4315e86..ff71e16cc752 100644 --- a/Documentation/cgroups/memory.txt +++ b/Documentation/cgroups/memory.txt @@ -493,6 +493,7 @@ pgpgin - # of charging events to the memory cgroup. The charging pgpgout - # of uncharging events to the memory cgroup. The uncharging event happens each time a page is unaccounted from the cgroup. swap - # of bytes of swap usage +dirty - # of bytes that are waiting to get written back to the disk. writeback - # of bytes of file/anon cache that are queued for syncing to disk. inactive_anon - # of bytes of anonymous and swap cache memory on inactive diff --git a/block/bio.c b/block/bio.c index 259197d97de1..2a00d349cd68 100644 --- a/block/bio.c +++ b/block/bio.c @@ -1988,6 +1988,28 @@ struct bio_set *bioset_create_nobvec(unsigned int pool_size, unsigned int front_ EXPORT_SYMBOL(bioset_create_nobvec); #ifdef CONFIG_BLK_CGROUP + +/** + * bio_associate_blkcg - associate a bio with the specified blkcg + * @bio: target bio + * @blkcg_css: css of the blkcg to associate + * + * Associate @bio with the blkcg specified by @blkcg_css. Block layer will + * treat @bio as if it were issued by a task which belongs to the blkcg. + * + * This function takes an extra reference of @blkcg_css which will be put + * when @bio is released. The caller must own @bio and is responsible for + * synchronizing calls to this function. + */ +int bio_associate_blkcg(struct bio *bio, struct cgroup_subsys_state *blkcg_css) +{ + if (unlikely(bio->bi_css)) + return -EBUSY; + css_get(blkcg_css); + bio->bi_css = blkcg_css; + return 0; +} + /** * bio_associate_current - associate a bio with %current * @bio: target bio @@ -2004,26 +2026,17 @@ EXPORT_SYMBOL(bioset_create_nobvec); int bio_associate_current(struct bio *bio) { struct io_context *ioc; - struct cgroup_subsys_state *css; - if (bio->bi_ioc) + if (bio->bi_css) return -EBUSY; ioc = current->io_context; if (!ioc) return -ENOENT; - /* acquire active ref on @ioc and associate */ get_io_context_active(ioc); bio->bi_ioc = ioc; - - /* associate blkcg if exists */ - rcu_read_lock(); - css = task_css(current, blkio_cgrp_id); - if (css && css_tryget_online(css)) - bio->bi_css = css; - rcu_read_unlock(); - + bio->bi_css = task_get_css(current, blkio_cgrp_id); return 0; } diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c index 6e43fa355e71..9f97da52d006 100644 --- a/block/blk-cgroup.c +++ b/block/blk-cgroup.c @@ -19,11 +19,12 @@ #include <linux/module.h> #include <linux/err.h> #include <linux/blkdev.h> +#include <linux/backing-dev.h> #include <linux/slab.h> #include <linux/genhd.h> #include <linux/delay.h> #include <linux/atomic.h> -#include "blk-cgroup.h" +#include <linux/blk-cgroup.h> #include "blk.h" #define MAX_KEY_LEN 100 @@ -33,6 +34,8 @@ static DEFINE_MUTEX(blkcg_pol_mutex); struct blkcg blkcg_root; EXPORT_SYMBOL_GPL(blkcg_root); +struct cgroup_subsys_state * const blkcg_root_css = &blkcg_root.css; + static struct blkcg_policy *blkcg_policy[BLKCG_MAX_POLS]; static bool blkcg_policy_enabled(struct request_queue *q, @@ -182,6 +185,7 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg, struct blkcg_gq *new_blkg) { struct blkcg_gq *blkg; + struct bdi_writeback_congested *wb_congested; int i, ret; WARN_ON_ONCE(!rcu_read_lock_held()); @@ -193,22 +197,30 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg, goto err_free_blkg; } + wb_congested = wb_congested_get_create(&q->backing_dev_info, + blkcg->css.id, GFP_ATOMIC); + if (!wb_congested) { + ret = -ENOMEM; + goto err_put_css; + } + /* allocate */ if (!new_blkg) { new_blkg = blkg_alloc(blkcg, q, GFP_ATOMIC); if (unlikely(!new_blkg)) { ret = -ENOMEM; - goto err_put_css; + goto err_put_congested; } } blkg = new_blkg; + blkg->wb_congested = wb_congested; /* link parent */ if (blkcg_parent(blkcg)) { blkg->parent = __blkg_lookup(blkcg_parent(blkcg), q, false); if (WARN_ON_ONCE(!blkg->parent)) { ret = -EINVAL; - goto err_put_css; + goto err_put_congested; } blkg_get(blkg->parent); } @@ -238,18 +250,15 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg, blkg->online = true; spin_unlock(&blkcg->lock); - if (!ret) { - if (blkcg == &blkcg_root) { - q->root_blkg = blkg; - q->root_rl.blkg = blkg; - } + if (!ret) return blkg; - } /* @blkg failed fully initialized, use the usual release path */ blkg_put(blkg); return ERR_PTR(ret); +err_put_congested: + wb_congested_put(wb_congested); err_put_css: css_put(&blkcg->css); err_free_blkg: @@ -343,15 +352,6 @@ static void blkg_destroy(struct blkcg_gq *blkg) rcu_assign_pointer(blkcg->blkg_hint, NULL); /* - * If root blkg is destroyed. Just clear the pointer since root_rl - * does not take reference on root blkg. - */ - if (blkcg == &blkcg_root) { - blkg->q->root_blkg = NULL; - blkg->q->root_rl.blkg = NULL; - } - - /* * Put the reference taken at the time of creation so that when all * queues are gone, group can be destroyed. */ @@ -405,6 +405,8 @@ void __blkg_release_rcu(struct rcu_head *rcu_head) if (blkg->parent) blkg_put(blkg->parent); + wb_congested_put(blkg->wb_congested); + blkg_free(blkg); } EXPORT_SYMBOL_GPL(__blkg_release_rcu); @@ -812,6 +814,8 @@ static void blkcg_css_offline(struct cgroup_subsys_state *css) } spin_unlock_irq(&blkcg->lock); + + wb_blkcg_offline(blkcg); } static void blkcg_css_free(struct cgroup_subsys_state *css) @@ -868,7 +872,9 @@ done: spin_lock_init(&blkcg->lock); INIT_RADIX_TREE(&blkcg->blkg_tree, GFP_ATOMIC); INIT_HLIST_HEAD(&blkcg->blkg_list); - +#ifdef CONFIG_CGROUP_WRITEBACK + INIT_LIST_HEAD(&blkcg->cgwb_list); +#endif return &blkcg->css; free_pd_blkcg: @@ -892,9 +898,45 @@ free_blkcg: */ int blkcg_init_queue(struct request_queue *q) { - might_sleep(); + struct blkcg_gq *new_blkg, *blkg; + bool preloaded; + int ret; + + new_blkg = blkg_alloc(&blkcg_root, q, GFP_KERNEL); + if (!new_blkg) + return -ENOMEM; + + preloaded = !radix_tree_preload(GFP_KERNEL); + + /* + * Make sure the root blkg exists and count the existing blkgs. As + * @q is bypassing at this point, blkg_lookup_create() can't be + * used. Open code insertion. + */ + rcu_read_lock(); + spin_lock_irq(q->queue_lock); + blkg = blkg_create(&blkcg_root, q, new_blkg); + spin_unlock_irq(q->queue_lock); + rcu_read_unlock(); + + if (preloaded) + radix_tree_preload_end(); + + if (IS_ERR(blkg)) { + kfree(new_blkg); + return PTR_ERR(blkg); + } - return blk_throtl_init(q); + q->root_blkg = blkg; + q->root_rl.blkg = blkg; + + ret = blk_throtl_init(q); + if (ret) { + spin_lock_irq(q->queue_lock); + blkg_destroy_all(q); + spin_unlock_irq(q->queue_lock); + } + return ret; } /** @@ -996,50 +1038,19 @@ int blkcg_activate_policy(struct request_queue *q, { LIST_HEAD(pds); LIST_HEAD(cpds); - struct blkcg_gq *blkg, *new_blkg; + struct blkcg_gq *blkg; struct blkg_policy_data *pd, *nd; struct blkcg_policy_data *cpd, *cnd; int cnt = 0, ret; - bool preloaded; if (blkcg_policy_enabled(q, pol)) return 0; - /* preallocations for root blkg */ - new_blkg = blkg_alloc(&blkcg_root, q, GFP_KERNEL); - if (!new_blkg) - return -ENOMEM; - + /* count and allocate policy_data for all existing blkgs */ blk_queue_bypass_start(q); - - preloaded = !radix_tree_preload(GFP_KERNEL); - - /* - * Make sure the root blkg exists and count the existing blkgs. As - * @q is bypassing at this point, blkg_lookup_create() can't be - * used. Open code it. - */ spin_lock_irq(q->queue_lock); - - rcu_read_lock(); - blkg = __blkg_lookup(&blkcg_root, q, false); - if (blkg) - blkg_free(new_blkg); - else - blkg = blkg_create(&blkcg_root, q, new_blkg); - rcu_read_unlock(); - - if (preloaded) - radix_tree_preload_end(); - - if (IS_ERR(blkg)) { - ret = PTR_ERR(blkg); - goto out_unlock; - } - list_for_each_entry(blkg, &q->blkg_list, q_node) cnt++; - spin_unlock_irq(q->queue_lock); /* @@ -1140,10 +1151,6 @@ void blkcg_deactivate_policy(struct request_queue *q, __clear_bit(pol->plid, q->blkcg_pols); - /* if no policy is left, no need for blkgs - shoot them down */ - if (bitmap_empty(q->blkcg_pols, BLKCG_MAX_POLS)) - blkg_destroy_all(q); - list_for_each_entry(blkg, &q->blkg_list, q_node) { /* grab blkcg lock too while removing @pd from @blkg */ spin_lock(&blkg->blkcg->lock); diff --git a/block/blk-core.c b/block/blk-core.c index f6ab750060fe..688ae9482cb8 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -32,12 +32,12 @@ #include <linux/delay.h> #include <linux/ratelimit.h> #include <linux/pm_runtime.h> +#include <linux/blk-cgroup.h> #define CREATE_TRACE_POINTS #include <trace/events/block.h> #include "blk.h" -#include "blk-cgroup.h" #include "blk-mq.h" EXPORT_TRACEPOINT_SYMBOL_GPL(block_bio_remap); @@ -63,6 +63,31 @@ struct kmem_cache *blk_requestq_cachep; */ static struct workqueue_struct *kblockd_workqueue; +static void blk_clear_congested(struct request_list *rl, int sync) +{ +#ifdef CONFIG_CGROUP_WRITEBACK + clear_wb_congested(rl->blkg->wb_congested, sync); +#else + /* + * If !CGROUP_WRITEBACK, all blkg's map to bdi->wb and we shouldn't + * flip its congestion state for events on other blkcgs. + */ + if (rl == &rl->q->root_rl) + clear_wb_congested(rl->q->backing_dev_info.wb.congested, sync); +#endif +} + +static void blk_set_congested(struct request_list *rl, int sync) +{ +#ifdef CONFIG_CGROUP_WRITEBACK + set_wb_congested(rl->blkg->wb_congested, sync); +#else + /* see blk_clear_congested() */ + if (rl == &rl->q->root_rl) + set_wb_congested(rl->q->backing_dev_info.wb.congested, sync); +#endif +} + void blk_queue_congestion_threshold(struct request_queue *q) { int nr; @@ -623,8 +648,7 @@ struct request_queue *blk_alloc_queue_node(gfp_t gfp_mask, int node_id) q->backing_dev_info.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE; - q->backing_dev_info.state = 0; - q->backing_dev_info.capabilities = 0; + q->backing_dev_info.capabilities = BDI_CAP_CGROUP_WRITEBACK; q->backing_dev_info.name = "block"; q->node = node_id; @@ -847,13 +871,8 @@ static void __freed_request(struct request_list *rl, int sync) { struct request_queue *q = rl->q; - /* - * bdi isn't aware of blkcg yet. As all async IOs end up root - * blkcg anyway, just use root blkcg state. - */ - if (rl == &q->root_rl && - rl->count[sync] < queue_congestion_off_threshold(q)) - blk_clear_queue_congested(q, sync); + if (rl->count[sync] < queue_congestion_off_threshold(q)) + blk_clear_congested(rl, sync); if (rl->count[sync] + 1 <= q->nr_requests) { if (waitqueue_active(&rl->wait[sync])) @@ -886,25 +905,25 @@ static void freed_request(struct request_list *rl, unsigned int flags) int blk_update_nr_requests(struct request_queue *q, unsigned int nr) { struct request_list *rl; + int on_thresh, off_thresh; spin_lock_irq(q->queue_lock); q->nr_requests = nr; blk_queue_congestion_threshold(q); + on_thresh = queue_congestion_on_threshold(q); + off_thresh = queue_congestion_off_threshold(q); - /* congestion isn't cgroup aware and follows root blkcg for now */ - rl = &q->root_rl; - - if (rl->count[BLK_RW_SYNC] >= queue_congestion_on_threshold(q)) - blk_set_queue_congested(q, BLK_RW_SYNC); - else if (rl->count[BLK_RW_SYNC] < queue_congestion_off_threshold(q)) - blk_clear_queue_congested(q, BLK_RW_SYNC); + blk_queue_for_each_rl(rl, q) { + if (rl->count[BLK_RW_SYNC] >= on_thresh) + blk_set_congested(rl, BLK_RW_SYNC); + else if (rl->count[BLK_RW_SYNC] < off_thresh) + blk_clear_congested(rl, BLK_RW_SYNC); - if (rl->count[BLK_RW_ASYNC] >= queue_congestion_on_threshold(q)) - blk_set_queue_congested(q, BLK_RW_ASYNC); - else if (rl->count[BLK_RW_ASYNC] < queue_congestion_off_threshold(q)) - blk_clear_queue_congested(q, BLK_RW_ASYNC); + if (rl->count[BLK_RW_ASYNC] >= on_thresh) + blk_set_congested(rl, BLK_RW_ASYNC); + else if (rl->count[BLK_RW_ASYNC] < off_thresh) + blk_clear_congested(rl, BLK_RW_ASYNC); - blk_queue_for_each_rl(rl, q) { if (rl->count[BLK_RW_SYNC] >= q->nr_requests) { blk_set_rl_full(rl, BLK_RW_SYNC); } else { @@ -1014,12 +1033,7 @@ static struct request *__get_request(struct request_list *rl, int rw_flags, } } } - /* - * bdi isn't aware of blkcg yet. As all async IOs end up - * root blkcg anyway, just use root blkcg state. - */ - if (rl == &q->root_rl) < |