// SPDX-License-Identifier: GPL-2.0 /* * Data Access Monitor * * Author: SeongJae Park <sj@kernel.org> */ #define pr_fmt(fmt) "damon: " fmt #include <linux/damon.h> #include <linux/delay.h> #include <linux/kthread.h> #include <linux/mm.h> #include <linux/psi.h> #include <linux/slab.h> #include <linux/string.h> #define CREATE_TRACE_POINTS #include <trace/events/damon.h> #ifdef CONFIG_DAMON_KUNIT_TEST #undef DAMON_MIN_REGION #define DAMON_MIN_REGION 1 #endif static DEFINE_MUTEX(damon_lock); static int nr_running_ctxs; static bool running_exclusive_ctxs; static DEFINE_MUTEX(damon_ops_lock); static struct damon_operations damon_registered_ops[NR_DAMON_OPS]; static struct kmem_cache *damon_region_cache __ro_after_init; /* Should be called under damon_ops_lock with id smaller than NR_DAMON_OPS */ static bool __damon_is_registered_ops(enum damon_ops_id id) { struct damon_operations empty_ops = {}; if (!memcmp(&empty_ops, &damon_registered_ops[id], sizeof(empty_ops))) return false; return true; } /** * damon_is_registered_ops() - Check if a given damon_operations is registered. * @id: Id of the damon_operations to check if registered. * * Return: true if the ops is set, false otherwise. */ bool damon_is_registered_ops(enum damon_ops_id id) { bool registered; if (id >= NR_DAMON_OPS) return false; mutex_lock(&damon_ops_lock); registered = __damon_is_registered_ops(id); mutex_unlock(&damon_ops_lock); return registered; } /** * damon_register_ops() - Register a monitoring operations set to DAMON. * @ops: monitoring operations set to register. * * This function registers a monitoring operations set of valid &struct * damon_operations->id so that others can find and use them later. * * Return: 0 on success, negative error code otherwise. */ int damon_register_ops(struct damon_operations *ops) { int err = 0; if (ops->id >= NR_DAMON_OPS) return -EINVAL; mutex_lock(&damon_ops_lock); /* Fail for already registered ops */ if (__damon_is_registered_ops(ops->id)) { err = -EINVAL; goto out; } damon_registered_ops[ops->id] = *ops; out: mutex_unlock(&damon_ops_lock); return err; } /** * damon_select_ops() - Select a monitoring operations to use with the context. * @ctx: monitoring context to use the operations. * @id: id of the registered monitoring operations to select. * * This function finds registered monitoring operations set of @id and make * @ctx to use it. * * Return: 0 on success, negative error code otherwise. */ int damon_select_ops(struct damon_ctx *ctx, enum damon_ops_id id) { int err = 0; if (id >= NR_DAMON_OPS) return -EINVAL; mutex_lock(&damon_ops_lock); if (!__damon_is_registered_ops(id)) err = -EINVAL; else ctx->ops = damon_registered_ops[id]; mutex_unlock(&damon_ops_lock); return err; } /* * Construct a damon_region struct * * Returns the pointer to the new struct if success, or NULL otherwise */ struct damon_region *damon_new_region(unsigned long start, unsigned long end) { struct damon_region *region; region = kmem_cache_alloc(damon_region_cache, GFP_KERNEL); if (!region) return NULL; region->ar.start = start; region->ar.end = end; region->nr_accesses = 0; region->nr_accesses_bp = 0; INIT_LIST_HEAD(®ion->list); region->age = 0; region->last_nr_accesses = 0; return region; } void damon_add_region(struct damon_region *r, struct damon_target *t) { list_add_tail(&r->list, &t->regions_list); t->nr_regions++; } static void damon_del_region(struct damon_region *r, struct damon_target *t) { list_del(&r->list); t->nr_regions--; } static void damon_free_region(struct damon_region *r) { kmem_cache_free(damon_region_cache, r); } void damon_destroy_region(struct damon_region *r, struct damon_target *t) { damon_del_region(r, t); damon_free_region(r); } /* * Check whether a region is intersecting an address range * * Returns true if it is. */ static bool damon_intersect(struct damon_region *r, struct damon_addr_range *re) { return !(r->ar.end <= re->start || re->end <= r->ar.start); } /* * Fill holes in regions with new regions. */ static int damon_fill_regions_holes(struct damon_region *first, struct damon_region *last, struct damon_target *t) { struct damon_region *r = first; damon_for_each_region_from(r, t) { struct damon_region *next, *newr; if (r == last) break; next = damon_next_region(r); if (r->ar.end != next->ar.start) { newr = damon_new_region(r->ar.end, next->ar.start); if (!newr) return -ENOMEM; damon_insert_region(newr, r, next, t); } } return 0; } /* * damon_set_regions() - Set regions of a target for given address ranges. * @t: the given target. * @ranges: array of new monitoring target ranges. * @nr_ranges: length of @ranges. * * This function adds new regions to, or modify existing regions of a * monitoring target to fit in specific ranges. * * Return: 0 if success, or negative error code otherwise. */ int damon_set_regions(struct damon_target *t, struct damon_addr_range *ranges, unsigned int nr_ranges) { struct damon_region *r, *next; unsigned int i; int err; /* Remove regions which are not in the new ranges */ damon_for_each_region_safe(r, next, t) { for (i = 0; i < nr_ranges; i++) { if (damon_intersect(r, &ranges[i])) break; } if (i == nr_ranges) damon_destroy_region(r, t); } r = damon_first_region(t); /* Add new regions or resize existing regions to fit in the ranges */ for (i = 0; i < nr_ranges; i++) { struct damon_region *first = NULL, *last, *newr; struct damon_addr_range *range; range = &ranges[i]; /* Get the first/last regions intersecting with the range */ damon_for_each_region_from(r, t) { if (damon_intersect(r, range)) { if (!first) first = r; last = r; } if (r->ar.start >= range->end) break; } if (!first) { /* no region intersects with this range */ newr = damon_new_region( ALIGN_DOWN(range->start, DAMON_MIN_REGION), ALIGN(range->end, DAMON_MIN_REGION)); if (!newr) return -ENOMEM; damon_insert_region(newr, damon_prev_region(r), r, t); } else { /* resize intersecting regions to fit in this range */ first->ar.start = ALIGN_DOWN(range->start, DAMON_MIN_REGION); last->ar.end = ALIGN(range->end, DAMON_MIN_REGION); /* fill possible holes in the range */ err = damon_fill_regions_holes(first, last, t); if (err) return err; } } return 0; } struct damos_filter *damos_new_filter(enum damos_filter_type type, bool matching) { struct damos_filter *filter; filter = kmalloc(sizeof(*filter), GFP_KERNEL); if (!filter) return NULL; filter->type = type; filter->matching = matching; INIT_LIST_HEAD(&filter->list); return filter; } void damos_add_filter(struct damos *s, struct damos_filter *f) { list_add_tail(&f->list, &s->filters); } static void damos_del_filter(struct damos_filter *f) { list_del(&f->list); } static void damos_free_filter(struct damos_filter *f) { kfree(f); } void damos_destroy_filter(struct damos_filter *f) { damos_del_filter(f); damos_free_filter(f); } struct damos_quota_goal *damos_new_quota_goal( enum damos_quota_goal_metric metric, unsigned long target_value) { struct damos_quota_goal *goal; goal = kmalloc(sizeof(*goal), GFP_KERNEL); if (!goal) return NULL; goal->metric = metric; goal->target_value = target_value; INIT_LIST_HEAD(&goal->list); return goal; } void damos_add_quota_goal(struct damos_quota *q, struct damos_quota_goal *g) { list_add_tail(&g->list, &q->goals); } static void damos_del_quota_goal(struct damos_quota_goal *g) { list_del(&g->list); } static void damos_free_quota_goal(struct damos_quota_goal *g) { kfree(g); } void damos_destroy_quota_goal(struct damos_quota_goal *g) { damos_del_quota_goal(g); damos_free_quota_goal(g); } /* initialize fields of @quota that normally API users wouldn't set */ static struct damos_quota *damos_quota_init(struct damos_quota *quota) { quota->esz = 0; quota->total_charged_sz = 0; quota->total_charged_ns = 0; quota->charged_sz = 0; quota->charged_from = 0; quota->charge_target_from = NULL; quota->charge_addr_from = 0; quota->esz_bp = 0; return quota; } struct damos *damon_new_scheme(struct damos_access_pattern *pattern, enum damos_action action, unsigned long apply_interval_us, struct damos_quota *quota, struct damos_watermarks *wmarks) { struct damos *scheme; scheme = kmalloc(sizeof(*scheme), GFP_KERNEL); if (!scheme) return NULL; scheme->pattern = *pattern; scheme->action = action; scheme->apply_interval_us = apply_interval_us; /* * next_apply_sis will be set when kdamond starts. While kdamond is * running, it will also updated when it is added to the DAMON context, * or damon_attrs are updated. */ scheme->next_apply_sis = 0; INIT_LIST_HEAD(&scheme->filters); scheme->stat = (struct damos_stat){}; INIT_LIST_HEAD(&scheme->list); scheme->quota = *(damos_quota_init(quota)); /* quota.goals should be separately set by caller */ INIT_LIST_HEAD(&scheme->quota.goals); scheme->wmarks = *wmarks; scheme->wmarks.activated = true; return scheme; } static void damos_set_next_apply_sis(struct damos *s, struct damon_ctx *ctx) { unsigned long sample_interval = ctx->attrs.sample_interval ? ctx->attrs.sample_interval : 1; unsigned long apply_interval = s->apply_interval_us ? s->apply_interval_us : ctx->attrs.aggr_interval; s->next_apply_sis = ctx->passed_sample_intervals + apply_interval / sample_interval; } void damon_add_scheme(struct damon_ctx *ctx, struct damos *s) { list_add_tail(&s->list, &ctx->schemes); damos_set_next_apply_sis(s, ctx); } static void damon_del_scheme(struct damos *s) { list_del(&s->list); } static void damon_free_scheme(struct damos *s) { kfree(s); } void damon_destroy_scheme(struct damos *s) { struct damos_quota_goal *g, *g_next; struct damos_filter *f, *next; damos_for_each_quota_goal_safe(g, g_next, &s->quota) damos_destroy_quota_goal(g); damos_for_each_filter_safe(f, next, s) damos_destroy_filter(f); damon_del_scheme(s); damon_free_scheme(s); } /* * Construct a damon_target struct * * Returns the pointer to the new struct if success, or NULL otherwise */ struct damon_target *damon_new_target(void) { struct damon_target *t; t = kmalloc(sizeof(*t), GFP_KERNEL); if (!t) return NULL; t->pid = NULL; t->nr_regions = 0; INIT_LIST_HEAD(&t->regions_list); INIT_LIST_HEAD(&t->list); return t; } void damon_add_target(struct damon_ctx *ctx, struct damon_target *t) { list_add_tail(&t->list, &ctx->adaptive_targets); } bool damon_targets_empty(struct damon_ctx *ctx) { return list_empty(&ctx->adaptive_targets); } static void damon_del_target(struct damon_target *t) { list_del(&t->list); } void damon_free_target(struct damon_target *t) { struct damon_region *r, *next; damon_for_each_region_safe(r, next, t) damon_free_region(r); kfree(t); } void damon_destroy_target(struct damon_target *t) { damon_del_target(t); damon_free_target(t); } unsigned int damon_nr_regions(struct damon_target *t) { return t->nr_regions; } struct damon_ctx *damon_new_ctx(void) { struct damon_ctx *ctx; ctx = kzalloc(sizeof(*ctx), GFP_KERNEL); if (!ctx) return NULL; init_completion(&ctx->kdamond_started); ctx->attrs.sample_interval = 5 * 1000; ctx->attrs.aggr_interval = 100 * 1000; ctx->attrs.ops_update_interval = 60 * 1000 * 1000; ctx->passed_sample_intervals = 0; /* These will be set from kdamond_init_intervals_sis() */ ctx->next_aggregation_sis = 0; ctx->next_ops_update_sis = 0; mutex_init(&ctx->kdamond_lock); ctx->attrs.min_nr_regions = 10; ctx->attrs.max_nr_regions = 1000; INIT_LIST_HEAD(&ctx->adaptive_targets); INIT_LIST_HEAD(&ctx->schemes); return ctx; } static void damon_destroy_targets(struct damon_ctx *ctx) { struct damon_target *t, *next_t; if (ctx->ops.cleanup) { ctx->ops.cleanup(ctx); return; } damon_for_each_target_safe(t, next_t, ctx) damon_destroy_target(t); } void damon_destroy_ctx(struct damon_ctx *ctx) { struct damos *s, *next_s; damon_destroy_targets(ctx); damon_for_each_scheme_safe(s, next_s, ctx) damon_destroy_scheme(s); kfree(ctx); } static unsigned int damon_age_for_new_attrs(unsigned int age, struct damon_attrs *old_attrs, struct damon_attrs *new_attrs) { return age * old_attrs->aggr_interval / new_attrs->aggr_interval; } /* convert access ratio in bp (per 10,000) to nr_accesses */ static unsigned int damon_accesses_bp_to_nr_accesses( unsigned int accesses_bp, struct damon_attrs *attrs) { return accesses_bp * damon_max_nr_accesses(attrs) / 10000; } /* convert nr_accesses to access ratio in bp (per 10,000) */ static unsigned int damon_nr_accesses_to_accesses_bp( unsigned int nr_accesses, struct damon_attrs *attrs) { return nr_accesses * 10000 / damon_max_nr_accesses(attrs); } static unsigned int damon_nr_accesses_for_new_attrs(unsigned int nr_accesses, struct damon_attrs *old_attrs, struct damon_attrs *new_attrs) { return damon_accesses_bp_to_nr_accesses( damon_nr_accesses_to_accesses_bp( nr_accesses, old_attrs), new_attrs); } static void damon_update_monitoring_result(struct damon_region *r, struct damon_attrs *old_attrs, struct damon_attrs *new_attrs) { r->nr_accesses = damon_nr_accesses_for_new_attrs(r->nr_accesses, old_attrs, new_attrs); r->nr_accesses_bp = r->nr_accesses * 10000; r->age = damon_age_for_new_attrs(r->age, old_attrs, new_attrs); } /* * region->nr_accesses is the number of sampling intervals in the last * aggregation interval that access to the region has found, and region->age is * the number of aggregation intervals that its access pattern has maintained. * For the reason, the real meaning of the two fields depend on current * sampling interval and aggregation interval. This function updates * ->nr_accesses and ->age of given damon_ctx's regions for new damon_attrs. */ static void damon_update_monitoring_results(struct damon_ctx *ctx, struct damon_attrs *new_attrs) { struct damon_attrs *old_attrs = &ctx->attrs; struct damon_target *t; struct damon_region *r; /* if any interval is zero, simply forgive conversion */ if (!old_attrs->sample_interval || !old_attrs->aggr_interval || !new_attrs->sample_interval || !new_attrs->aggr_interval) return; damon_for_each_target(t, ctx) damon_for_each_region(r, t) damon_update_monitoring_result( r, old_attrs, new_attrs); } /** * damon_set_attrs() - Set attributes for the monitoring. * @ctx: monitoring context * @attrs: monitoring attributes * * This function should be called while the kdamond is not running, or an * access check results aggregation is not ongoing (e.g., from * &struct damon_callback->after_aggregation or * &struct damon_callback->after_wmarks_check callbacks). * * Every time interval is in micro-seconds. * * Return: 0 on success, negative error code otherwise. */ int damon_set_attrs(struct damon_ctx *ctx, struct damon_attrs *attrs) { unsigned long sample_interval = attrs->sample_interval ? attrs->sample_interval : 1; struct damos *s; if (attrs->min_nr_regions < 3) return -EINVAL; if (attrs->min_nr_regions > attrs->max_nr_regions) return -EINVAL; if (attrs->sample_interval > attrs->aggr_interval) return -EINVAL; ctx->next_aggregation_sis = ctx->passed_sample_intervals + attrs->aggr_interval / sample_interval; ctx->next_ops_update_sis = ctx->passed_sample_intervals + attrs->ops_update_interval / sample_interval; damon_update_monitoring_results(ctx, attrs); ctx->attrs = *attrs; damon_for_each_scheme(s, ctx) damos_set_next_apply_sis(s, ctx); return 0; } /** * damon_set_schemes() - Set data access monitoring based operation schemes. * @ctx: monitoring context * @schemes: array of the schemes * @nr_schemes: number of entries in @schemes * * This function should not be called while the kdamond of the context is * running. */ void damon_set_schemes(struct damon_ctx *ctx, struct damos **schemes, ssize_t nr_schemes) { struct damos *s, *next; ssize_t i; damon_for_each_scheme_safe(s, next, ctx) damon_destroy_scheme(s); for (i = 0; i < nr_schemes; i++) damon_add_scheme(ctx, schemes[i]); } /** * damon_nr_running_ctxs() - Return number of currently running contexts. */ int damon_nr_running_ctxs(void) { int nr_ctxs; mutex_lock(&damon_lock); nr_ctxs = nr_running_ctxs; mutex_unlock(&damon_lock); return nr_ctxs; } /* Returns the size upper limit for each monitoring region */ static unsigned long damon_region_sz_limit(struct damon_ctx *ctx) { struct damon_target *t; struct damon_region *r; unsigned long sz = 0; damon_for_each_target(t, ctx) { damon_for_each_region(r, t) sz += damon_sz_region(r); } if (ctx->attrs.min_nr_regions) sz /= ctx->attrs.min_nr_regions; if (sz < DAMON_MIN_REGION) sz = DAMON_MIN_REGION; return sz; } static int kdamond_fn(void *data); /* * __damon_start() - Starts monitoring with given context. * @ctx: monitoring context * * This function should be called while damon_lock is hold. * * Return: 0 on success, negative error code otherwise. */ static int __damon_start(struct damon_ctx *ctx) { int err = -EBUSY; mutex_lock(&ctx->kdamond_lock); if (!ctx->kdamond) { err = 0; reinit_completion(&ctx->kdamond_started); ctx->kdamond = kthread_run(kdamond_fn, ctx, "kdamond.%d", nr_running_ctxs); if (IS_ERR(ctx->kdamond)) { err = PTR_ERR(ctx->kdamond); ctx->kdamond = NULL; } else { wait_for_completion(&ctx->kdamond_started); } } mutex_unlock(&ctx->kdamond_lock); return err; } /** * damon_start() - Starts the monitorings for a given group of contexts. * @ctxs: an array of the pointers for contexts to start monitoring * @nr_ctxs: size of @ctxs * @exclusive: exclusiveness of this contexts group * * This function starts a group of monitoring threads for a group of monitoring * contexts. One thread per each context is created and run in parallel. The * caller should handle synchronization between the threads by itself. If * @exclusive is true and a group of threads that created by other * 'damon_start()' call is currently running, this function does nothing but * returns -EBUSY. * * Return: 0 on success, negative error code otherwise. */ int damon_start(struct damon_ctx **ctxs, int nr_ctxs, bool exclusive) { int i; int err = 0; mutex_lock(&damon_lock); if ((exclusive && nr_running_ctxs) || (!exclusive && running_exclusive_ctxs)) { mutex_unlock(&damon_lock); return -EBUSY; } for (i = 0; i < nr_ctxs; i++) { err = __damon_start(ctxs[i]); if (err) break; nr_running_ctxs++; } if (exclusive && nr_running_ctxs) running_exclusive_ctxs = true; mutex_unlock(&damon_lock); return err; } /* * __damon_stop() - Stops monitoring of a given context. * @ctx: monitoring context * * Return: 0 on success, negative error code otherwise. */ static int __damon_stop(struct damon_ctx *ctx) { struct task_struct *tsk; mutex_lock(&ctx->kdamond_lock); tsk = ctx->kdamond; if (tsk) { get_task_struct(tsk); mutex_unlock(&ctx->kdamond_lock); kthread_stop_put(tsk); return 0; } mutex_unlock(&ctx->kdamond_lock); return -EPERM; } /** * damon_stop() - Stops the monitorings for a given group of contexts. * @ctxs: an array of the pointers for contexts to stop monitoring * @nr_ctxs: size of @ctxs * * Return: 0 on success, negative error code otherwise. */ int damon_stop(struct damon_ctx **ctxs, int nr_ctxs) { int i, err = 0; for (i = 0; i < nr_ctxs; i++) { /* nr_running_ctxs is decremented in kdamond_fn */ err = __damon_stop(ctxs[i]); if (err) break; } return err; } /* * Reset the aggregated monitoring results ('nr_accesses' of each region). */ static void kdamond_reset_aggregated(struct damon_ctx *c) { struct damon_target *t; unsigned int ti = 0; /* target's index */ damon_for_each_target(t, c) { struct damon_region *r; damon_for_each_region(r, t) { trace_damon_aggregated(ti, r, damon_nr_regions(t)); r->last_nr_accesses = r->nr_accesses; r->nr_accesses = 0; } ti++; } } static void damon_split_region_at(struct damon_target *t, struct damon_region *r, unsigned long sz_r); static bool __damos_valid_target(struct damon_region *r, struct damos *s) { unsigned long sz; unsigned int nr_accesses = r->nr_accesses_bp / 10000; sz = damon_sz_region(r); return s->pattern.min_sz_region <= sz && sz <= s->pattern.max_sz_region && s->pattern.min_nr_accesses <= nr_accesses && nr_accesses <= s->pattern.max_nr_accesses && s->pattern.min_age_region <= r->age && r->age <= s->pattern.max_age_region; } static bool damos_valid_target(struct damon_ctx *c, struct damon_target *t, struct damon_region *r, struct damos *s) { bool ret = __damos_valid_target(r, s); if (!ret || !s->quota.esz || !c->ops.get_scheme_score) return ret; return c->ops.get_scheme_score(c, t, r, s) >= s->quota.min_score; } /* * damos_skip_charged_region() - Check if the given region or starting part of * it is already charged for the DAMOS quota. * @t: The target of the region. * @rp: The pointer to the region. * @s: The scheme to be applied. * * If a quota of a scheme has exceeded in a quota charge window, the scheme's * action would applied to only a part of the target access pattern fulfilling * regions. To avoid applying the scheme action to only already applied * regions, DAMON skips applying the scheme action to the regions that charged * in the previous charge window. * * This function checks if a given region should be skipped or not for the * reason. If only the starting part of the region has previously charged, * this function splits the region into two so that the second one covers the * area that not charged in the previous charge widnow and saves the second * region in *rp and returns false, so that the caller can apply DAMON action * to the second one. * * Return: true if the region should be entirely skipped, false otherwise. */ static bool damos_skip_charged_region(struct damon_target *t, struct damon_region **rp, struct damos *s) { struct damon_region *r = *rp; struct damos_quota *quota = &s->quota; unsigned long sz_to_skip; /* Skip previously charged regions */ if (quota->charge_target_from) { if (t != quota->charge_target_from) return true; if (r == damon_last_region(t)) { quota->charge_target_from = NULL; quota->charge_addr_from = 0; return true; } if (quota->charge_addr_from && r->ar.end <= quota->charge_addr_from) return true; if (quota->charge_addr_from && r->ar.start < quota->charge_addr_from) { sz_to_skip = ALIGN_DOWN(quota->charge_addr_from - r->ar.start, DAMON_MIN_REGION); if (!sz_to_skip) { if (damon_sz_region(r) <= DAMON_MIN_REGION) return true; sz_to_skip = DAMON_MIN_REGION; } damon_split_region_at(t, r, sz_to_skip); r = damon_next_region(r); *rp = r; } quota->charge_target_from = NULL; quota->charge_addr_from = 0; } return false; } static void damos_update_stat(struct damos *s, unsigned long sz_tried, unsigned long sz_applied) { s->stat.nr_tried++; s->stat.sz_tried += sz_tried; if (sz_applied) s->stat.nr_applied++; s->stat.sz_applied += sz_applied; } static bool __damos_filter_out(struct damon_ctx *ctx, struct damon_target *t, struct damon_region *r, struct damos_filter *filter) { bool matched = false; struct damon_target *ti; int target_idx = 0; unsigned long start, end; switch (filter->type) { case DAMOS_FILTER_TYPE_TARGET: damon_for_each_target(ti, ctx) { if (ti == t) break; target_idx++; } matched = target_idx == filter->target_idx; break; case DAMOS_FILTER_TYPE_ADDR: start = ALIGN_DOWN(filter->addr_range.start, DAMON_MIN_REGION); end = ALIGN_DOWN(filter->addr_range.end, DAMON_MIN_REGION); /* inside the range */ if (start <= r->ar.start && r->ar.end <= end) { matched = true; break; } /* outside of the range */ if (r->ar.end <= start || end <= r->ar.start) { matched = false; break; } /* start before the range and overlap */ if (r->ar.start < start) { damon_split_region_at(t, r, start - r->ar.start); matched = false; break; } /* start inside the range */ damon_split_region_at(t, r, end - r->ar.start); matched = true; break; default: return false; } return matched == filter->matching; } static bool damos_filter_out(struct damon_ctx *ctx, struct damon_target *t, struct damon_region *r, struct damos *s) { struct damos_filter *filter; damos_for_each_filter(filter, s) { if (__damos_filter_out(ctx, t, r, filter)) return true; } return false; } static void damos_apply_scheme(struct damon_ctx *c, struct damon_target *t, struct damon_region *r, struct damos *s) { struct damos_quota *quota = &s->quota; unsigned long sz = damon_sz_region(r); struct timespec64 begin, end; unsigned long sz_applied = 0; int err = 0; /* * We plan to support multiple context per kdamond, as DAMON sysfs * implies with 'nr_contexts' file. Nevertheless, only single context * per kdamond is supported for now. So, we can simply use '0' context * index here. */ unsigned int cidx = 0; struct damos *siter; /* schemes iterator */ unsigned int sidx = 0; struct damon_target *titer; /* targets iterator */ unsigned int tidx = 0; bool do_trace = false; /* get indices for trace_damos_before_apply() */ if (trace_damos_before_apply_enabled()) { damon_for_each_scheme(siter, c) { if (siter == s) break; sidx++; } damon_for_each_target(titer, c) { if (titer == t) break; tidx++; } do_trace = true; } if (c->ops.apply_scheme) { if (quota->esz && quota->charged_sz + sz > quota->esz) { sz = ALIGN_DOWN(quota->esz - quota->charged_sz, DAMON_MIN_REGION); if (!sz) goto update_stat; damon_split_region_at(t, r, sz); } if (damos_filter_out(c, t, r, s)) return; ktime_get_coarse_ts64(&begin); if (c->callback.before_damos_apply) err = c->callback.before_damos_apply(c, t, r, s); if (!err) { trace_damos_before_apply(cidx, sidx, tidx, r, damon_nr_regions(t), do_trace); sz_applied = c->ops.apply_scheme(c, t, r, s); } ktime_get_coarse_ts64(&end); quota->total_charged_ns += timespec64_to_ns(&end) - timespec64_to_ns(&begin); quota->charged_sz += sz; if (quota->esz && quota->charged_sz >= quota->esz) { quota->charge_target_from = t; quota->charge_addr_from = r->ar.end + 1; } } if (s->action != DAMOS_STAT) r->age = 0; update_stat: damos_update_stat(s, sz, sz_applied); } static void damon_do_apply_schemes(struct damon_ctx *c, struct damon_target *t, struct damon_region *r) { struct damos *s; damon_for_each_scheme(s, c) { struct damos_quota *quota = &s->quota; if (c->passed_sample_intervals != s->next_apply_sis) continue; if (!s->wmarks.activated) continue; /* Check the quota */ if (quota->esz && quota->charged_sz >= quota->esz) continue; if (damos_skip_charged_region(t, &r, s)) continue; if (!damos_valid_target(c, t, r, s)) continue; damos_apply_scheme(c, t, r, s); } } /* * damon_feed_loop_next_input() - get next input to achieve a target score. * @last_input The last input. * @score Current score that made with @last_input. * * Calculate next input to achieve the target score, based on the last input * and current score. Assuming the input and the score are positively * proportional, calculate how much compensation should be added to or * subtracted from the last input as a proportion of the last input. Avoid * next input always being zero by setting it non-zero always. In short form * (assuming support of float and signed calculations), the algorithm is as * below. * * next_input = max(last_input * ((goal - current) / goal + 1), 1) * * For simple implementation, we assume the target score is always 10,000. The * caller should adjust @score for this. * * Returns next input that assumed to achieve the target score. */ static unsigned long damon_feed_loop_next_input(unsigned long last_input, unsigned long score) { const unsigned long goal = 10000; unsigned long score_goal_diff = max(goal, score) - min(goal, score); unsigned long score_goal_diff_bp = score_goal_diff * 10000 / goal; unsigned long compensation = last_input * score_goal_diff_bp / 10000; /* Set minimum input as 10000 to avoid compensation be zero */ const unsigned long min_input = 10000; if (goal > score) return last_input + compensation; if (last_input > compensation + min_input) return last_input - compensation; return min_input; } #ifdef CONFIG_PSI static u64 damos_get_some_mem_psi_total(void) { if (static_branch_likely(&psi_disabled)) return 0; return div_u64(psi_system.total[PSI_AVGS][PSI_MEM * 2], NSEC_PER_USEC); } #else /* CONFIG_PSI */ static inline u64 damos_get_some_mem_psi_total(void) { return 0; }; #endif /* CONFIG_PSI */ static void damos_set_quota_goal_current_value(struct damos_quota_goal *goal) { u64 now_psi_total; switch (goal->metric) { case DAMOS_QUOTA_USER_INPUT: /* User should already set goal->current_value */ break; case DAMOS_QUOTA_SOME_MEM_PSI_US: now_psi_total = damos_get_some_mem_psi_total(); goal->current_value = now_psi_total - goal->last_psi_total; goal->last_psi_total = now_psi_total; break; default: break; } } /* Return the highest score since it makes schemes least aggressive */ static unsigned long damos_quota_score(struct damos_quota *quota) { struct damos_quota_goal *goal; unsigned long highest_score = 0; damos_for_each_quota_goal(goal, quota) { damos_set_quota_goal_current_value(goal); highest_score = max(highest_score, goal->current_value * 10000 / goal->target_value); } return highest_score; } /* * Called only if quota->ms, or quota->sz are set, or quota->goals is not empty */ static void damos_set_effective_quota(struct damos_quota *quota) { unsigned long throughput; unsigned long esz; if (!quota->ms && list_empty("a->goals)) { quota->esz = quota->sz; return; } if (!list_empty("a->goals)) { unsigned long score = damos_quota_score(quota); quota->esz_bp = damon_feed_loop_next_input( max(quota->esz_bp, 10000UL), score); esz = quota->esz_bp / 10000; } if (quota->ms) { if (quota->total_charged_ns) throughput = quota->total_charged_sz * 1000000 / quota->total_charged_ns; else throughput = PAGE_SIZE * 1024; if (!list_empty("a->goals)) esz = min(throughput * quota->ms, esz); else esz = throughput * quota->ms; } if (quota->sz && quota->sz < esz) esz = quota->sz; quota->esz = esz; } static void damos_adjust_quota(struct damon_ctx *c, struct damos *s) { struct damos_quota *quota = &s->quota; struct damon_target *t; struct damon_region *r; unsigned long cumulated_sz; unsigned int score, max_score = 0; if (!quota->ms && !quota->sz && list_empty("a->goals)) return; /* New charge window starts */ if (time_after_eq(jiffies, quota->charged_from + msecs_to_jiffies(quota->reset_interval))) { if (quota->esz && quota->charged_sz >= quota->esz) s->stat.qt_exceeds++; quota->total_charged_sz += quota->charged_sz; quota->charged_from = jiffies; quota->charged_sz = 0; damos_set_effective_quota(quota); } if (!c->ops.get_scheme_score) return; /* Fill up the score histogram */ memset(quota->histogram, 0, sizeof(quota->histogram)); damon_for_each_target(t, c) { damon_for_each_region(r, t) { if (!__damos_valid_target(r, s)) continue; score = c->ops.get_scheme_score(c, t, r, s); quota->histogram[score] += damon_sz_region(r); if (score > max_score) max_score = score; } } /* Set the min score limit */ for (cumulated_sz = 0, score = max_score; ; score--) { cumulated_sz += quota->histogram[score]; if (cumulated_sz >= quota->esz || !score) break; } quota->min_score = score; } static void kdamond_apply_schemes(struct damon_ctx *c) { struct damon_target *t; struct damon_region *r, *next_r; struct damos *s; unsigned long sample_interval = c->attrs.sample_interval ? c->attrs.sample_interval : 1; bool has_schemes_to_apply = false; damon_for_each_scheme(s, c) { if (c->passed_sample_intervals != s->next_apply_sis) continue; if (!s->wmarks.activated) continue; has_schemes_to_apply = true; damos_adjust_quota(c, s); } if (!has_schemes_to_apply) return; damon_for_each_target(t, c) { damon_for_each_region_safe(r, next_r, t) damon_do_apply_schemes(c, t, r); } damon_for_each_scheme(s, c) { if (c->passed_sample_intervals != s->next_apply_sis) continue; s->next_apply_sis += (s->apply_interval_us ? s->apply_interval_us : c->attrs.aggr_interval) / sample_interval; } } /* * Merge two adjacent regions into one region */ static void damon_merge_two_regions(struct damon_target *t, struct damon_region *l, struct damon_region *r) { unsigned long sz_l = damon_sz_region(l), sz_r = damon_sz_region(r); l->nr_accesses = (l->nr_accesses * sz_l + r->nr_accesses * sz_r) / (sz_l + sz_r); l->nr_accesses_bp = l->nr_accesses * 10000; l->age = (l->age * sz_l + r->age * sz_r) / (sz_l + sz_r); l->ar.end = r->ar.end; damon_destroy_region(r, t); } /* * Merge adjacent regions having similar access frequencies * * t target affected by this merge operation * thres '->nr_accesses' diff threshold for the merge * sz_limit size upper limit of each region */ static void damon_merge_regions_of(struct damon_target *t, unsigned int thres, unsigned long sz_limit) { struct damon_region *r, *prev = NULL, *next; damon_for_each_region_safe(r, next, t) { if (abs(r->nr_accesses - r->last_nr_accesses) > thres) r->age = 0; else r->age++; if (prev && prev->ar.end == r->ar.start && abs(prev->nr_accesses - r->nr_accesses) <= thres && damon_sz_region(prev) + damon_sz_region(r) <= sz_limit) damon_merge_two_regions(t, prev, r); else prev = r; } } /* * Merge adjacent regions having similar access frequencies * * threshold '->nr_accesses' diff threshold for the merge * sz_limit size upper limit of each region * * This function merges monitoring target regions which are adjacent and their * access frequencies are similar. This is for minimizing the monitoring * overhead under the dynamically changeable access pattern. If a merge was * unnecessarily made, later 'kdamond_split_regions()' will revert it. */ static void kdamond_merge_regions(struct damon_ctx *c, unsigned int threshold, unsigned long sz_limit) { struct damon_target *t; damon_for_each_target(t, c) damon_merge_regions_of(t, threshold, sz_limit); } /* * Split a region in two * * r the region to be split * sz_r size of the first sub-region that will be made */ static void damon_split_region_at(struct damon_target *t, struct damon_region *r, unsigned long sz_r) { struct damon_region *new; new = damon_new_region(r->ar.start + sz_r, r->ar.end); if (!new) return; r->ar.end = new->ar.start; new->age = r->age; new->last_nr_accesses = r->last_nr_accesses; new->nr_accesses_bp = r->nr_accesses_bp; new->nr_accesses = r->nr_accesses; damon_insert_region(new, r, damon_next_region(r), t); } /* Split every region in the given target into 'nr_subs' regions */ static void damon_split_regions_of(struct damon_target *t, int nr_subs) { struct damon_region *r, *next; unsigned long sz_region, sz_sub = 0; int i; damon_for_each_region_safe(r, next, t) { sz_region = damon_sz_region(r); for (i = 0; i < nr_subs - 1 && sz_region > 2 * DAMON_MIN_REGION; i++) { /* * Randomly select size of left sub-region to be at * least 10 percent and at most 90% of original region */ sz_sub = ALIGN_DOWN(damon_rand(1, 10) * sz_region / 10, DAMON_MIN_REGION); /* Do not allow blank region */ if (sz_sub == 0 || sz_sub >= sz_region) continue; damon_split_region_at(t, r, sz_sub); sz_region = sz_sub; } } } /* * Split every target region into randomly-sized small regions * * This function splits every target region into random-sized small regions if * current total number of the regions is equal or smaller than half of the * user-specified maximum number of regions. This is for maximizing the * monitoring accuracy under the dynamically changeable access patterns. If a * split was unnecessarily made, later 'kdamond_merge_regions()' will revert * it. */ static void kdamond_split_regions(struct damon_ctx *ctx) { struct damon_target *t; unsigned int nr_regions = 0; static unsigned int last_nr_regions; int nr_subregions = 2; damon_for_each_target(t, ctx) nr_regions += damon_nr_regions(t); if (nr_regions > ctx->attrs.max_nr_regions / 2) return; /* Maybe the middle of the region has different access frequency */ if (last_nr_regions == nr_regions && nr_regions < ctx->attrs.max_nr_regions / 3) nr_subregions = 3; damon_for_each_target(t, ctx) damon_split_regions_of(t, nr_subregions); last_nr_regions = nr_regions; } /* * Check whether current monitoring should be stopped * * The monitoring is stopped when either the user requested to stop, or all * monitoring targets are invalid. * * Returns true if need to stop current monitoring. */ static bool kdamond_need_stop(struct damon_ctx *ctx) { struct damon_target *t; if (kthread_should_stop()) return true; if (!ctx->ops.target_valid) return false; damon_for_each_target(t, ctx) { if (ctx->ops.target_valid(t)) return false; } return true; } static int damos_get_wmark_metric_value(enum damos_wmark_metric metric, unsigned long *metric_value) { switch (metric) { case DAMOS_WMARK_FREE_MEM_RATE: *metric_value = global_zone_page_state(NR_FREE_PAGES) * 1000 / totalram_pages(); return 0; default: break; } return -EINVAL; } /* * Returns zero if the scheme is active. Else, returns time to wait for next * watermark check in micro-seconds. */ static unsigned long damos_wmark_wait_us(struct damos *scheme) { unsigned long metric; if (damos_get_wmark_metric_value(scheme->wmarks.metric, &metric)) return 0; /* higher than high watermark or lower than low watermark */ if (metric > scheme->wmarks.high || scheme->wmarks.low > metric) { if (scheme->wmarks.activated) pr_debug("deactivate a scheme (%d) for %s wmark\n", scheme->action, metric > scheme->wmarks.high ? "high" : "low"); scheme->wmarks.activated = false; return scheme->wmarks.interval; } /* inactive and higher than middle watermark */ if ((scheme->wmarks.high >= metric && metric >= scheme->wmarks.mid) && !scheme->wmarks.activated) return scheme->wmarks.interval; if (!scheme->wmarks.activated) pr_debug("activate a scheme (%d)\n", scheme->action); scheme->wmarks.activated = true; return 0; } static void kdamond_usleep(unsigned long usecs) { /* See Documentation/timers/timers-howto.rst for the thresholds */ if (usecs > 20 * USEC_PER_MSEC) schedule_timeout_idle(usecs_to_jiffies(usecs)); else usleep_idle_range(usecs, usecs + 1); } /* Returns negative error code if it's not activated but should return */ static int kdamond_wait_activation(struct damon_ctx *ctx) { struct damos *s; unsigned long wait_time; unsigned long min_wait_time = 0; bool init_wait_time = false; while (!kdamond_need_stop(ctx)) { damon_for_each_scheme(s, ctx) { wait_time = damos_wmark_wait_us(s); if (!init_wait_time || wait_time < min_wait_time) { init_wait_time = true; min_wait_time = wait_time; } } if (!min_wait_time) return 0; kdamond_usleep(min_wait_time); if (ctx->callback.after_wmarks_check && ctx->callback.after_wmarks_check(ctx)) break; } return -EBUSY; } static void kdamond_init_intervals_sis(struct damon_ctx *ctx) { unsigned long sample_interval = ctx->attrs.sample_interval ? ctx->attrs.sample_interval : 1; unsigned long apply_interval; struct damos *scheme; ctx->passed_sample_intervals = 0; ctx->next_aggregation_sis = ctx->attrs.aggr_interval / sample_interval; ctx->next_ops_update_sis = ctx->attrs.ops_update_interval / sample_interval; damon_for_each_scheme(scheme, ctx) { apply_interval = scheme->apply_interval_us ? scheme->apply_interval_us : ctx->attrs.aggr_interval; scheme->next_apply_sis = apply_interval / sample_interval; } } /* * The monitoring daemon that runs as a kernel thread */ static int kdamond_fn(void *data) { struct damon_ctx *ctx = data; struct damon_target *t; struct damon_region *r, *next; unsigned int max_nr_accesses = 0; unsigned long sz_limit = 0; pr_debug("kdamond (%d) starts\n", current->pid); complete(&ctx->kdamond_started); kdamond_init_intervals_sis(ctx); if (ctx->ops.init) ctx->ops.init(ctx); if (ctx->callback.before_start && ctx->callback.before_start(ctx)) goto done; sz_limit = damon_region_sz_limit(ctx); while (!kdamond_need_stop(ctx)) { /* * ctx->attrs and ctx->next_{aggregation,ops_update}_sis could * be changed from after_wmarks_check() or after_aggregation() * callbacks. Read the values here, and use those for this * iteration. That is, damon_set_attrs() updated new values * are respected from next iteration. */ unsigned long next_aggregation_sis = ctx->next_aggregation_sis; unsigned long next_ops_update_sis = ctx->next_ops_update_sis; unsigned long sample_interval = ctx->attrs.sample_interval; if (kdamond_wait_activation(ctx)) break; if (ctx->ops.prepare_access_checks) ctx->ops.prepare_access_checks(ctx); if (ctx->callback.after_sampling && ctx->callback.after_sampling(ctx)) break; kdamond_usleep(sample_interval); ctx->passed_sample_intervals++; if (ctx->ops.check_accesses) max_nr_accesses = ctx->ops.check_accesses(ctx); if (ctx->passed_sample_intervals == next_aggregation_sis) { kdamond_merge_regions(ctx, max_nr_accesses / 10, sz_limit); if (ctx->callback.after_aggregation && ctx->callback.after_aggregation(ctx)) break; } /* * do kdamond_apply_schemes() after kdamond_merge_regions() if * possible, to reduce overhead */ if (!list_empty(&ctx->schemes)) kdamond_apply_schemes(ctx); sample_interval = ctx->attrs.sample_interval ? ctx->attrs.sample_interval : 1; if (ctx->passed_sample_intervals == next_aggregation_sis) { ctx->next_aggregation_sis = next_aggregation_sis + ctx->attrs.aggr_interval / sample_interval; kdamond_reset_aggregated(ctx); kdamond_split_regions(ctx); if (ctx->ops.reset_aggregated) ctx->ops.reset_aggregated(ctx); } if (ctx->passed_sample_intervals == next_ops_update_sis) { ctx->next_ops_update_sis = next_ops_update_sis + ctx->attrs.ops_update_interval / sample_interval; if (ctx->ops.update) ctx->ops.update(ctx); sz_limit = damon_region_sz_limit(ctx); } } done: damon_for_each_target(t, ctx) { damon_for_each_region_safe(r, next, t) damon_destroy_region(r, t); } if (ctx->callback.before_terminate) ctx->callback.before_terminate(ctx); if (ctx->ops.cleanup) ctx->ops.cleanup(ctx); pr_debug("kdamond (%d) finishes\n", current->pid); mutex_lock(&ctx->kdamond_lock); ctx->kdamond = NULL; mutex_unlock(&ctx->kdamond_lock); mutex_lock(&damon_lock); nr_running_ctxs--; if (!nr_running_ctxs && running_exclusive_ctxs) running_exclusive_ctxs = false; mutex_unlock(&damon_lock); return 0; } /* * struct damon_system_ram_region - System RAM resource address region of * [@start, @end). * @start: Start address of the region (inclusive). * @end: End address of the region (exclusive). */ struct damon_system_ram_region { unsigned long start; unsigned long end; }; static int walk_system_ram(struct resource *res, void *arg) { struct damon_system_ram_region *a = arg; if (a->end - a->start < resource_size(res)) { a->start = res->start; a->end = res->end; } return 0; } /* * Find biggest 'System RAM' resource and store its start and end address in * @start and @end, respectively. If no System RAM is found, returns false. */ static bool damon_find_biggest_system_ram(unsigned long *start, unsigned long *end) { struct damon_system_ram_region arg = {}; walk_system_ram_res(0, ULONG_MAX, &arg, walk_system_ram); if (arg.end <= arg.start) return false; *start = arg.start; *end = arg.end; return true; } /** * damon_set_region_biggest_system_ram_default() - Set the region of the given * monitoring target as requested, or biggest 'System RAM'. * @t: The monitoring target to set the region. * @start: The pointer to the start address of the region. * @end: The pointer to the end address of the region. * * This function sets the region of @t as requested by @start and @end. If the * values of @start and @end are zero, however, this function finds the biggest * 'System RAM' resource and sets the region to cover the resource. In the * latter case, this function saves the start and end addresses of the resource * in @start and @end, respectively. * * Return: 0 on success, negative error code otherwise. */ int damon_set_region_biggest_system_ram_default(struct damon_target *t, unsigned long *start, unsigned long *end) { struct damon_addr_range addr_range; if (*start > *end) return -EINVAL; if (!*start && !*end && !damon_find_biggest_system_ram(start, end)) return -EINVAL; addr_range.start = *start; addr_range.end = *end; return damon_set_regions(t, &addr_range, 1); } /* * damon_moving_sum() - Calculate an inferred moving sum value. * @mvsum: Inferred sum of the last @len_window values. * @nomvsum: Non-moving sum of the last discrete @len_window window values. * @len_window: The number of last values to take care of. * @new_value: New value that will be added to the pseudo moving sum. * * Moving sum (moving average * window size) is good for handling noise, but * the cost of keeping past values can be high for arbitrary window size. This * function implements a lightweight pseudo moving sum function that doesn't * keep the past window values. * * It simply assumes there was no noise in the past, and get the no-noise * assumed past value to drop from @nomvsum and @len_window. @nomvsum is a * non-moving sum of the last window. For example, if @len_window is 10 and we * have 25 values, @nomvsum is the sum of the 11th to 20th values of the 25 * values. Hence, this function simply drops @nomvsum / @len_window from * given @mvsum and add @new_value. * * For example, if @len_window is 10 and @nomvsum is 50, the last 10 values for * the last window could be vary, e.g., 0, 10, 0, 10, 0, 10, 0, 0, 0, 20. For * calculating next moving sum with a new value, we should drop 0 from 50 and * add the new value. However, this function assumes it got value 5 for each * of the last ten times. Based on the assumption, when the next value is * measured, it drops the assumed past value, 5 from the current sum, and add * the new value to get the updated pseduo-moving average. * * This means the value could have errors, but the errors will be disappeared * for every @len_window aligned calls. For example, if @len_window is 10, the * pseudo moving sum with 11th value to 19th value would have an error. But * the sum with 20th value will not have the error. * * Return: Pseudo-moving average after getting the @new_value. */ static unsigned int damon_moving_sum(unsigned int mvsum, unsigned int nomvsum, unsigned int len_window, unsigned int new_value) { return mvsum - nomvsum / len_window + new_value; } /** * damon_update_region_access_rate() - Update the access rate of a region. * @r: The DAMON region to update for its access check result. * @accessed: Whether the region has accessed during last sampling interval. * @attrs: The damon_attrs of the DAMON context. * * Update the access rate of a region with the region's last sampling interval * access check result. * * Usually this will be called by &damon_operations->check_accesses callback. */ void damon_update_region_access_rate(struct damon_region *r, bool accessed, struct damon_attrs *attrs) { unsigned int len_window = 1; /* * sample_interval can be zero, but cannot be larger than * aggr_interval, owing to validation of damon_set_attrs(). */ if (attrs->sample_interval) len_window = damon_max_nr_accesses(attrs); r->nr_accesses_bp = damon_moving_sum(r->nr_accesses_bp, r->last_nr_accesses * 10000, len_window, accessed ? 10000 : 0); if (accessed) r->nr_accesses++; } static int __init damon_init(void) { damon_region_cache = KMEM_CACHE(damon_region, 0); if (unlikely(!damon_region_cache)) { pr_err("creating damon_region_cache fails\n"); return -ENOMEM; } return 0; } subsys_initcall(damon_init); #include "core-test.h"