1 /* 2 * CDDL HEADER START 3 * 4 * This file and its contents are supplied under the terms of the 5 * Common Development and Distribution License ("CDDL"), version 1.0. 6 * You may only use this file in accordance with the terms of version 7 * 1.0 of the CDDL. 8 * 9 * A full copy of the text of the CDDL should have accompanied this 10 * source. A copy of the CDDL is also available via the Internet at 11 * http://www.illumos.org/license/CDDL. 12 * 13 * CDDL HEADER END 14 */ 15 /* 16 * Copyright (c) 2017 by Delphix. All rights reserved. 17 */ 18 19 #include <sys/zfs_context.h> 20 #include <sys/aggsum.h> 21 22 /* 23 * Aggregate-sum counters are a form of fanned-out counter, used when atomic 24 * instructions on a single field cause enough CPU cache line contention to 25 * slow system performance. Due to their increased overhead and the expense 26 * involved with precisely reading from them, they should only be used in cases 27 * where the write rate (increment/decrement) is much higher than the read rate 28 * (get value). 29 * 30 * Aggregate sum counters are comprised of two basic parts, the core and the 31 * buckets. The core counter contains a lock for the entire counter, as well 32 * as the current upper and lower bounds on the value of the counter. The 33 * aggsum_bucket structure contains a per-bucket lock to protect the contents of 34 * the bucket, the current amount that this bucket has changed from the global 35 * counter (called the delta), and the amount of increment and decrement we have 36 * "borrowed" from the core counter. 37 * 38 * The basic operation of an aggsum is simple. Threads that wish to modify the 39 * counter will modify one bucket's counter (determined by their current CPU, to 40 * help minimize lock and cache contention). If the bucket already has 41 * sufficient capacity borrowed from the core structure to handle their request, 42 * they simply modify the delta and return. If the bucket does not, we clear 43 * the bucket's current state (to prevent the borrowed amounts from getting too 44 * large), and borrow more from the core counter. Borrowing is done by adding to 45 * the upper bound (or subtracting from the lower bound) of the core counter, 46 * and setting the borrow value for the bucket to the amount added (or 47 * subtracted). Clearing the bucket is the opposite; we add the current delta 48 * to both the lower and upper bounds of the core counter, subtract the borrowed 49 * incremental from the upper bound, and add the borrowed decrement from the 50 * lower bound. Note that only borrowing and clearing require access to the 51 * core counter; since all other operations access CPU-local resources, 52 * performance can be much higher than a traditional counter. 53 * 54 * Threads that wish to read from the counter have a slightly more challenging 55 * task. It is fast to determine the upper and lower bounds of the aggum; this 56 * does not require grabbing any locks. This suffices for cases where an 57 * approximation of the aggsum's value is acceptable. However, if one needs to 58 * know whether some specific value is above or below the current value in the 59 * aggsum, they invoke aggsum_compare(). This function operates by repeatedly 60 * comparing the target value to the upper and lower bounds of the aggsum, and 61 * then clearing a bucket. This proceeds until the target is outside of the 62 * upper and lower bounds and we return a response, or the last bucket has been 63 * cleared and we know that the target is equal to the aggsum's value. Finally, 64 * the most expensive operation is determining the precise value of the aggsum. 65 * To do this, we clear every bucket and then return the upper bound (which must 66 * be equal to the lower bound). What makes aggsum_compare() and aggsum_value() 67 * expensive is clearing buckets. This involves grabbing the global lock 68 * (serializing against themselves and borrow operations), grabbing a bucket's 69 * lock (preventing threads on those CPUs from modifying their delta), and 70 * zeroing out the borrowed value (forcing that thread to borrow on its next 71 * request, which will also be expensive). This is what makes aggsums well 72 * suited for write-many read-rarely operations. 73 */ 74 75 /* 76 * We will borrow aggsum_borrow_multiplier times the current request, so we will 77 * have to get the as_lock approximately every aggsum_borrow_multiplier calls to 78 * aggsum_delta(). 79 */ 80 static uint_t aggsum_borrow_multiplier = 10; 81 82 void 83 aggsum_init(aggsum_t *as, uint64_t value) 84 { 85 bzero(as, sizeof (*as)); 86 as->as_lower_bound = as->as_upper_bound = value; 87 mutex_init(&as->as_lock, NULL, MUTEX_DEFAULT, NULL); 88 as->as_numbuckets = boot_ncpus; 89 as->as_buckets = kmem_zalloc(boot_ncpus * sizeof (aggsum_bucket_t), 90 KM_SLEEP); 91 for (int i = 0; i < as->as_numbuckets; i++) { 92 mutex_init(&as->as_buckets[i].asc_lock, 93 NULL, MUTEX_DEFAULT, NULL); 94 } 95 } 96 97 void 98 aggsum_fini(aggsum_t *as) 99 { 100 for (int i = 0; i < as->as_numbuckets; i++) 101 mutex_destroy(&as->as_buckets[i].asc_lock); 102 mutex_destroy(&as->as_lock); 103 } 104 105 int64_t 106 aggsum_lower_bound(aggsum_t *as) 107 { 108 return (as->as_lower_bound); 109 } 110 111 int64_t 112 aggsum_upper_bound(aggsum_t *as) 113 { 114 return (as->as_upper_bound); 115 } 116 117 static void 118 aggsum_flush_bucket(aggsum_t *as, struct aggsum_bucket *asb) 119 { 120 ASSERT(MUTEX_HELD(&as->as_lock)); 121 ASSERT(MUTEX_HELD(&asb->asc_lock)); 122 123 /* 124 * We use atomic instructions for this because we read the upper and 125 * lower bounds without the lock, so we need stores to be atomic. 126 */ 127 atomic_add_64((volatile uint64_t *)&as->as_lower_bound, asb->asc_delta); 128 atomic_add_64((volatile uint64_t *)&as->as_upper_bound, asb->asc_delta); 129 asb->asc_delta = 0; 130 atomic_add_64((volatile uint64_t *)&as->as_upper_bound, 131 -asb->asc_borrowed); 132 atomic_add_64((volatile uint64_t *)&as->as_lower_bound, 133 asb->asc_borrowed); 134 asb->asc_borrowed = 0; 135 } 136 137 uint64_t 138 aggsum_value(aggsum_t *as) 139 { 140 int64_t rv; 141 142 mutex_enter(&as->as_lock); 143 if (as->as_lower_bound == as->as_upper_bound) { 144 rv = as->as_lower_bound; 145 for (int i = 0; i < as->as_numbuckets; i++) { 146 ASSERT0(as->as_buckets[i].asc_delta); 147 ASSERT0(as->as_buckets[i].asc_borrowed); 148 } 149 mutex_exit(&as->as_lock); 150 return (rv); 151 } 152 for (int i = 0; i < as->as_numbuckets; i++) { 153 struct aggsum_bucket *asb = &as->as_buckets[i]; 154 mutex_enter(&asb->asc_lock); 155 aggsum_flush_bucket(as, asb); 156 mutex_exit(&asb->asc_lock); 157 } 158 VERIFY3U(as->as_lower_bound, ==, as->as_upper_bound); 159 rv = as->as_lower_bound; 160 mutex_exit(&as->as_lock); 161 162 return (rv); 163 } 164 165 static void 166 aggsum_borrow(aggsum_t *as, int64_t delta, struct aggsum_bucket *asb) 167 { 168 int64_t abs_delta = (delta < 0 ? -delta : delta); 169 mutex_enter(&as->as_lock); 170 mutex_enter(&asb->asc_lock); 171 172 aggsum_flush_bucket(as, asb); 173 174 atomic_add_64((volatile uint64_t *)&as->as_upper_bound, abs_delta); 175 atomic_add_64((volatile uint64_t *)&as->as_lower_bound, -abs_delta); 176 asb->asc_borrowed = abs_delta; 177 178 mutex_exit(&asb->asc_lock); 179 mutex_exit(&as->as_lock); 180 } 181 182 void 183 aggsum_add(aggsum_t *as, int64_t delta) 184 { 185 struct aggsum_bucket *asb = 186 &as->as_buckets[CPU_SEQID % as->as_numbuckets]; 187 188 for (;;) { 189 mutex_enter(&asb->asc_lock); 190 if (asb->asc_delta + delta <= (int64_t)asb->asc_borrowed && 191 asb->asc_delta + delta >= -(int64_t)asb->asc_borrowed) { 192 asb->asc_delta += delta; 193 mutex_exit(&asb->asc_lock); 194 return; 195 } 196 mutex_exit(&asb->asc_lock); 197 aggsum_borrow(as, delta * aggsum_borrow_multiplier, asb); 198 } 199 } 200 201 /* 202 * Compare the aggsum value to target efficiently. Returns -1 if the value 203 * represented by the aggsum is less than target, 1 if it's greater, and 0 if 204 * they are equal. 205 */ 206 int 207 aggsum_compare(aggsum_t *as, uint64_t target) 208 { 209 if (as->as_upper_bound < target) 210 return (-1); 211 if (as->as_lower_bound > target) 212 return (1); 213 mutex_enter(&as->as_lock); 214 for (int i = 0; i < as->as_numbuckets; i++) { 215 struct aggsum_bucket *asb = &as->as_buckets[i]; 216 mutex_enter(&asb->asc_lock); 217 aggsum_flush_bucket(as, asb); 218 mutex_exit(&asb->asc_lock); 219 if (as->as_upper_bound < target) { 220 mutex_exit(&as->as_lock); 221 return (-1); 222 } 223 if (as->as_lower_bound > target) { 224 mutex_exit(&as->as_lock); 225 return (1); 226 } 227 } 228 VERIFY3U(as->as_lower_bound, ==, as->as_upper_bound); 229 ASSERT3U(as->as_lower_bound, ==, target); 230 mutex_exit(&as->as_lock); 231 return (0); 232 } 233