^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0-only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * mm_init.c - Memory initialisation verification and debugging
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * Copyright 2008 IBM Corporation, 2008
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) * Author Mel Gorman <mel@csn.ul.ie>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include <linux/kernel.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include <linux/init.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include <linux/kobject.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include <linux/export.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include <linux/memory.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <linux/notifier.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <linux/sched.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include <linux/mman.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include "internal.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #ifdef CONFIG_DEBUG_MEMORY_INIT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) int __meminitdata mminit_loglevel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #ifndef SECTIONS_SHIFT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #define SECTIONS_SHIFT 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) /* The zonelists are simply reported, validation is manual. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) void __init mminit_verify_zonelist(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) int nid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) if (mminit_loglevel < MMINIT_VERIFY)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) for_each_online_node(nid) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) pg_data_t *pgdat = NODE_DATA(nid);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) struct zone *zone;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) struct zoneref *z;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) struct zonelist *zonelist;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) int i, listid, zoneid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) BUILD_BUG_ON(MAX_ZONELISTS > 2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) for (i = 0; i < MAX_ZONELISTS * MAX_NR_ZONES; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) /* Identify the zone and nodelist */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) zoneid = i % MAX_NR_ZONES;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) listid = i / MAX_NR_ZONES;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) zonelist = &pgdat->node_zonelists[listid];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) zone = &pgdat->node_zones[zoneid];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) if (!populated_zone(zone))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) /* Print information about the zonelist */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) printk(KERN_DEBUG "mminit::zonelist %s %d:%s = ",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) listid > 0 ? "thisnode" : "general", nid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) zone->name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) /* Iterate the zonelist */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) for_each_zone_zonelist(zone, z, zonelist, zoneid)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) pr_cont("%d:%s ", zone_to_nid(zone), zone->name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) pr_cont("\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) void __init mminit_verify_pageflags_layout(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) int shift, width;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) unsigned long or_mask, add_mask;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) shift = 8 * sizeof(unsigned long);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) width = shift - SECTIONS_WIDTH - NODES_WIDTH - ZONES_WIDTH
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) - LAST_CPUPID_SHIFT - KASAN_TAG_WIDTH;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) mminit_dprintk(MMINIT_TRACE, "pageflags_layout_widths",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) "Section %d Node %d Zone %d Lastcpupid %d Kasantag %d Flags %d\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) SECTIONS_WIDTH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) NODES_WIDTH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) ZONES_WIDTH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) LAST_CPUPID_WIDTH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) KASAN_TAG_WIDTH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) NR_PAGEFLAGS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) mminit_dprintk(MMINIT_TRACE, "pageflags_layout_shifts",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) "Section %d Node %d Zone %d Lastcpupid %d Kasantag %d\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) SECTIONS_SHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) NODES_SHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) ZONES_SHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) LAST_CPUPID_SHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) KASAN_TAG_WIDTH);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) mminit_dprintk(MMINIT_TRACE, "pageflags_layout_pgshifts",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) "Section %lu Node %lu Zone %lu Lastcpupid %lu Kasantag %lu\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) (unsigned long)SECTIONS_PGSHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) (unsigned long)NODES_PGSHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) (unsigned long)ZONES_PGSHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) (unsigned long)LAST_CPUPID_PGSHIFT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) (unsigned long)KASAN_TAG_PGSHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) mminit_dprintk(MMINIT_TRACE, "pageflags_layout_nodezoneid",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) "Node/Zone ID: %lu -> %lu\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) (unsigned long)(ZONEID_PGOFF + ZONEID_SHIFT),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) (unsigned long)ZONEID_PGOFF);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) mminit_dprintk(MMINIT_TRACE, "pageflags_layout_usage",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) "location: %d -> %d layout %d -> %d unused %d -> %d page-flags\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) shift, width, width, NR_PAGEFLAGS, NR_PAGEFLAGS, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) #ifdef NODE_NOT_IN_PAGE_FLAGS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) mminit_dprintk(MMINIT_TRACE, "pageflags_layout_nodeflags",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) "Node not in page flags");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) #ifdef LAST_CPUPID_NOT_IN_PAGE_FLAGS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) mminit_dprintk(MMINIT_TRACE, "pageflags_layout_nodeflags",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) "Last cpupid not in page flags");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) if (SECTIONS_WIDTH) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) shift -= SECTIONS_WIDTH;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) BUG_ON(shift != SECTIONS_PGSHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) if (NODES_WIDTH) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) shift -= NODES_WIDTH;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) BUG_ON(shift != NODES_PGSHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) if (ZONES_WIDTH) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) shift -= ZONES_WIDTH;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) BUG_ON(shift != ZONES_PGSHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) /* Check for bitmask overlaps */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) or_mask = (ZONES_MASK << ZONES_PGSHIFT) |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) (NODES_MASK << NODES_PGSHIFT) |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) (SECTIONS_MASK << SECTIONS_PGSHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) add_mask = (ZONES_MASK << ZONES_PGSHIFT) +
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) (NODES_MASK << NODES_PGSHIFT) +
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) (SECTIONS_MASK << SECTIONS_PGSHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) BUG_ON(or_mask != add_mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) static __init int set_mminit_loglevel(char *str)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) get_option(&str, &mminit_loglevel);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) early_param("mminit_loglevel", set_mminit_loglevel);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) #endif /* CONFIG_DEBUG_MEMORY_INIT */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) struct kobject *mm_kobj;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) EXPORT_SYMBOL_GPL(mm_kobj);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) #ifdef CONFIG_SMP
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) s32 vm_committed_as_batch = 32;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) void mm_compute_batch(int overcommit_policy)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) u64 memsized_batch;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) s32 nr = num_present_cpus();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) s32 batch = max_t(s32, nr*2, 32);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) unsigned long ram_pages = totalram_pages();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) * For policy OVERCOMMIT_NEVER, set batch size to 0.4% of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) * (total memory/#cpus), and lift it to 25% for other policies
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) * to easy the possible lock contention for percpu_counter
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) * vm_committed_as, while the max limit is INT_MAX
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) if (overcommit_policy == OVERCOMMIT_NEVER)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) memsized_batch = min_t(u64, ram_pages/nr/256, INT_MAX);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) memsized_batch = min_t(u64, ram_pages/nr/4, INT_MAX);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) vm_committed_as_batch = max_t(s32, memsized_batch, batch);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) static int __meminit mm_compute_batch_notifier(struct notifier_block *self,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) unsigned long action, void *arg)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) switch (action) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) case MEM_ONLINE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) case MEM_OFFLINE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) mm_compute_batch(sysctl_overcommit_memory);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) return NOTIFY_OK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) static struct notifier_block compute_batch_nb __meminitdata = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) .notifier_call = mm_compute_batch_notifier,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) .priority = IPC_CALLBACK_PRI, /* use lowest priority */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) static int __init mm_compute_batch_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) mm_compute_batch(sysctl_overcommit_memory);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) register_hotmemory_notifier(&compute_batch_nb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) __initcall(mm_compute_batch_init);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) static int __init mm_sysfs_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) mm_kobj = kobject_create_and_add("mm", kernel_kobj);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) if (!mm_kobj)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) postcore_initcall(mm_sysfs_init);