^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) * Copyright (C) 1993 Linus Torvalds
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * SMP-safe vmalloc/vfree/ioremap, Tigran Aivazian <tigran@veritas.com>, May 2000
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) * Major rework to support vmap/vunmap, Christoph Hellwig, SGI, August 2002
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * Numa awareness, Christoph Lameter, SGI, June 2005
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) * Improving global KVA allocator, Uladzislau Rezki, Sony, May 2019
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include <linux/vmalloc.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include <linux/mm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include <linux/module.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <linux/highmem.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <linux/sched/signal.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include <linux/slab.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include <linux/spinlock.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include <linux/interrupt.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include <linux/proc_fs.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include <linux/seq_file.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include <linux/set_memory.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include <linux/debugobjects.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #include <linux/kallsyms.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #include <linux/list.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) #include <linux/notifier.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) #include <linux/rbtree.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) #include <linux/xarray.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) #include <linux/rcupdate.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #include <linux/pfn.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) #include <linux/kmemleak.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) #include <linux/atomic.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) #include <linux/compiler.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) #include <linux/llist.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) #include <linux/bitops.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) #include <linux/rbtree_augmented.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) #include <linux/overflow.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) #include <trace/hooks/mm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) #include <linux/uaccess.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) #include <asm/tlbflush.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) #include <asm/shmparam.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) #include "internal.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) #include "pgalloc-track.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) bool is_vmalloc_addr(const void *x)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) unsigned long addr = (unsigned long)x;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) return addr >= VMALLOC_START && addr < VMALLOC_END;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) EXPORT_SYMBOL(is_vmalloc_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) struct vfree_deferred {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) struct llist_head list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) struct work_struct wq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) static DEFINE_PER_CPU(struct vfree_deferred, vfree_deferred);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) static void __vunmap(const void *, int);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) static void free_work(struct work_struct *w)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) struct vfree_deferred *p = container_of(w, struct vfree_deferred, wq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) struct llist_node *t, *llnode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) llist_for_each_safe(llnode, t, llist_del_all(&p->list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) __vunmap((void *)llnode, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) /*** Page table manipulation functions ***/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) static void vunmap_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) pgtbl_mod_mask *mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) pte_t *pte;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) pte = pte_offset_kernel(pmd, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) pte_t ptent = ptep_get_and_clear(&init_mm, addr, pte);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) WARN_ON(!pte_none(ptent) && !pte_present(ptent));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) } while (pte++, addr += PAGE_SIZE, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) *mask |= PGTBL_PTE_MODIFIED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) static void vunmap_pmd_range(pud_t *pud, unsigned long addr, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) pgtbl_mod_mask *mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) pmd_t *pmd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) int cleared;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) pmd = pmd_offset(pud, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) next = pmd_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) cleared = pmd_clear_huge(pmd);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) if (cleared || pmd_bad(*pmd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) *mask |= PGTBL_PMD_MODIFIED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) if (cleared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) if (pmd_none_or_clear_bad(pmd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) vunmap_pte_range(pmd, addr, next, mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) } while (pmd++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) static void vunmap_pud_range(p4d_t *p4d, unsigned long addr, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) pgtbl_mod_mask *mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) pud_t *pud;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) int cleared;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) pud = pud_offset(p4d, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) next = pud_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) cleared = pud_clear_huge(pud);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) if (cleared || pud_bad(*pud))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) *mask |= PGTBL_PUD_MODIFIED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) if (cleared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) if (pud_none_or_clear_bad(pud))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) vunmap_pmd_range(pud, addr, next, mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) } while (pud++, addr = next, addr != end);
^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 void vunmap_p4d_range(pgd_t *pgd, unsigned long addr, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) pgtbl_mod_mask *mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) p4d_t *p4d;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) int cleared;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) p4d = p4d_offset(pgd, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) next = p4d_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) cleared = p4d_clear_huge(p4d);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) if (cleared || p4d_bad(*p4d))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) *mask |= PGTBL_P4D_MODIFIED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) if (cleared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) if (p4d_none_or_clear_bad(p4d))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) vunmap_pud_range(p4d, addr, next, mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) } while (p4d++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) * unmap_kernel_range_noflush - unmap kernel VM area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) * @start: start of the VM area to unmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) * @size: size of the VM area to unmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) * Unmap PFN_UP(@size) pages at @addr. The VM area @addr and @size specify
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) * should have been allocated using get_vm_area() and its friends.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) * NOTE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) * This function does NOT do any cache flushing. The caller is responsible
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) * for calling flush_cache_vunmap() on to-be-mapped areas before calling this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) * function and flush_tlb_kernel_range() after.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) void unmap_kernel_range_noflush(unsigned long start, unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) unsigned long end = start + size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) pgd_t *pgd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) unsigned long addr = start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) pgtbl_mod_mask mask = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) BUG_ON(addr >= end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) pgd = pgd_offset_k(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) next = pgd_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) if (pgd_bad(*pgd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) mask |= PGTBL_PGD_MODIFIED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) if (pgd_none_or_clear_bad(pgd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) vunmap_p4d_range(pgd, addr, next, &mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) } while (pgd++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) if (mask & ARCH_PAGE_TABLE_SYNC_MASK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) arch_sync_kernel_mappings(start, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) static int vmap_pte_range(pmd_t *pmd, unsigned long addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) unsigned long end, pgprot_t prot, struct page **pages, int *nr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) pgtbl_mod_mask *mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) pte_t *pte;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) * nr is a running index into the array which helps higher level
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) * callers keep track of where we're up to.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) pte = pte_alloc_kernel_track(pmd, addr, mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) if (!pte)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) struct page *page = pages[*nr];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) if (WARN_ON(!pte_none(*pte)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) return -EBUSY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) if (WARN_ON(!page))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) set_pte_at(&init_mm, addr, pte, mk_pte(page, prot));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) (*nr)++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) } while (pte++, addr += PAGE_SIZE, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) *mask |= PGTBL_PTE_MODIFIED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) static int vmap_pmd_range(pud_t *pud, unsigned long addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) unsigned long end, pgprot_t prot, struct page **pages, int *nr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) pgtbl_mod_mask *mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) pmd_t *pmd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) pmd = pmd_alloc_track(&init_mm, pud, addr, mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) if (!pmd)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) next = pmd_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) if (vmap_pte_range(pmd, addr, next, prot, pages, nr, mask))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) } while (pmd++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) static int vmap_pud_range(p4d_t *p4d, unsigned long addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) unsigned long end, pgprot_t prot, struct page **pages, int *nr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) pgtbl_mod_mask *mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) pud_t *pud;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) pud = pud_alloc_track(&init_mm, p4d, addr, mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) if (!pud)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) next = pud_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) if (vmap_pmd_range(pud, addr, next, prot, pages, nr, mask))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) } while (pud++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) static int vmap_p4d_range(pgd_t *pgd, unsigned long addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) unsigned long end, pgprot_t prot, struct page **pages, int *nr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) pgtbl_mod_mask *mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) p4d_t *p4d;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) p4d = p4d_alloc_track(&init_mm, pgd, addr, mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) if (!p4d)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) next = p4d_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) if (vmap_pud_range(p4d, addr, next, prot, pages, nr, mask))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) } while (p4d++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) * map_kernel_range_noflush - map kernel VM area with the specified pages
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) * @addr: start of the VM area to map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) * @size: size of the VM area to map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) * @prot: page protection flags to use
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) * @pages: pages to map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) * Map PFN_UP(@size) pages at @addr. The VM area @addr and @size specify should
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) * have been allocated using get_vm_area() and its friends.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) * NOTE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) * This function does NOT do any cache flushing. The caller is responsible for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) * calling flush_cache_vmap() on to-be-mapped areas before calling this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) * function.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) * RETURNS:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) * 0 on success, -errno on failure.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) int map_kernel_range_noflush(unsigned long addr, unsigned long size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) pgprot_t prot, struct page **pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) unsigned long start = addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) unsigned long end = addr + size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) pgd_t *pgd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) int err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) int nr = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) pgtbl_mod_mask mask = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) BUG_ON(addr >= end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) pgd = pgd_offset_k(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) next = pgd_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) if (pgd_bad(*pgd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) mask |= PGTBL_PGD_MODIFIED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) err = vmap_p4d_range(pgd, addr, next, prot, pages, &nr, &mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) if (err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) } while (pgd++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) if (mask & ARCH_PAGE_TABLE_SYNC_MASK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) arch_sync_kernel_mappings(start, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) int map_kernel_range(unsigned long start, unsigned long size, pgprot_t prot,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) struct page **pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) ret = map_kernel_range_noflush(start, size, prot, pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) flush_cache_vmap(start, start + size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) EXPORT_SYMBOL_GPL(map_kernel_range);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) int is_vmalloc_or_module_addr(const void *x)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) * ARM, x86-64 and sparc64 put modules in a special place,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) * and fall back on vmalloc() if that fails. Others
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) * just put it in the vmalloc space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) #if defined(CONFIG_MODULES) && defined(MODULES_VADDR)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) unsigned long addr = (unsigned long)x;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) if (addr >= MODULES_VADDR && addr < MODULES_END)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) return is_vmalloc_addr(x);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) * Walk a vmap address to the struct page it maps.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) struct page *vmalloc_to_page(const void *vmalloc_addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) unsigned long addr = (unsigned long) vmalloc_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) struct page *page = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) pgd_t *pgd = pgd_offset_k(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) p4d_t *p4d;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) pud_t *pud;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) pmd_t *pmd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) pte_t *ptep, pte;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) * XXX we might need to change this if we add VIRTUAL_BUG_ON for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) * architectures that do not vmalloc module space
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) VIRTUAL_BUG_ON(!is_vmalloc_or_module_addr(vmalloc_addr));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) if (pgd_none(*pgd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) p4d = p4d_offset(pgd, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) if (p4d_none(*p4d))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) pud = pud_offset(p4d, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) * Don't dereference bad PUD or PMD (below) entries. This will also
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) * identify huge mappings, which we may encounter on architectures
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) * that define CONFIG_HAVE_ARCH_HUGE_VMAP=y. Such regions will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) * identified as vmalloc addresses by is_vmalloc_addr(), but are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) * not [unambiguously] associated with a struct page, so there is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) * no correct value to return for them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) WARN_ON_ONCE(pud_bad(*pud));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) if (pud_none(*pud) || pud_bad(*pud))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) pmd = pmd_offset(pud, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) WARN_ON_ONCE(pmd_bad(*pmd));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) if (pmd_none(*pmd) || pmd_bad(*pmd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) ptep = pte_offset_map(pmd, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) pte = *ptep;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) if (pte_present(pte))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) page = pte_page(pte);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) pte_unmap(ptep);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) return page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) EXPORT_SYMBOL(vmalloc_to_page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) * Map a vmalloc()-space virtual address to the physical page frame number.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) unsigned long vmalloc_to_pfn(const void *vmalloc_addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) return page_to_pfn(vmalloc_to_page(vmalloc_addr));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) EXPORT_SYMBOL(vmalloc_to_pfn);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) /*** Global kva allocator ***/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) #define DEBUG_AUGMENT_PROPAGATE_CHECK 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) #define DEBUG_AUGMENT_LOWEST_MATCH_CHECK 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) static DEFINE_SPINLOCK(vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) static DEFINE_SPINLOCK(free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) /* Export for kexec only */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) LIST_HEAD(vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) static LLIST_HEAD(vmap_purge_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) static struct rb_root vmap_area_root = RB_ROOT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) static bool vmap_initialized __read_mostly;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) * This kmem_cache is used for vmap_area objects. Instead of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) * allocating from slab we reuse an object from this cache to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) * make things faster. Especially in "no edge" splitting of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) * free block.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) static struct kmem_cache *vmap_area_cachep;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) * This linked list is used in pair with free_vmap_area_root.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) * It gives O(1) access to prev/next to perform fast coalescing.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) static LIST_HEAD(free_vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) * This augment red-black tree represents the free vmap space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) * All vmap_area objects in this tree are sorted by va->va_start
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) * address. It is used for allocation and merging when a vmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) * object is released.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) * Each vmap_area node contains a maximum available free block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) * of its sub-tree, right or left. Therefore it is possible to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) * find a lowest match of free area.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) static struct rb_root free_vmap_area_root = RB_ROOT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) * Preload a CPU with one object for "no edge" split case. The
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) * aim is to get rid of allocations from the atomic context, thus
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) * to use more permissive allocation masks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) static DEFINE_PER_CPU(struct vmap_area *, ne_fit_preload_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) static __always_inline unsigned long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) va_size(struct vmap_area *va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) return (va->va_end - va->va_start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) static __always_inline unsigned long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) get_subtree_max_size(struct rb_node *node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) va = rb_entry_safe(node, struct vmap_area, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) return va ? va->subtree_max_size : 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) * Gets called when remove the node and rotate.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) static __always_inline unsigned long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) compute_subtree_max_size(struct vmap_area *va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) return max3(va_size(va),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) get_subtree_max_size(va->rb_node.rb_left),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) get_subtree_max_size(va->rb_node.rb_right));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) RB_DECLARE_CALLBACKS_MAX(static, free_vmap_area_rb_augment_cb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) struct vmap_area, rb_node, unsigned long, subtree_max_size, va_size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) static void purge_vmap_area_lazy(void);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) static BLOCKING_NOTIFIER_HEAD(vmap_notify_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) static unsigned long lazy_max_pages(void);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) static atomic_long_t nr_vmalloc_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) unsigned long vmalloc_nr_pages(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) return atomic_long_read(&nr_vmalloc_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) EXPORT_SYMBOL_GPL(vmalloc_nr_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) static struct vmap_area *__find_vmap_area(unsigned long addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) struct rb_node *n = vmap_area_root.rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) while (n) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) va = rb_entry(n, struct vmap_area, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) if (addr < va->va_start)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) n = n->rb_left;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) else if (addr >= va->va_end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) n = n->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) return va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) * This function returns back addresses of parent node
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) * and its left or right link for further processing.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) * Otherwise NULL is returned. In that case all further
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) * steps regarding inserting of conflicting overlap range
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) * have to be declined and actually considered as a bug.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) static __always_inline struct rb_node **
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) find_va_links(struct vmap_area *va,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) struct rb_root *root, struct rb_node *from,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) struct rb_node **parent)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) struct vmap_area *tmp_va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) struct rb_node **link;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) if (root) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) link = &root->rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) if (unlikely(!*link)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) *parent = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) return link;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) link = &from;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) * Go to the bottom of the tree. When we hit the last point
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) * we end up with parent rb_node and correct direction, i name
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) * it link, where the new va->rb_node will be attached to.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) tmp_va = rb_entry(*link, struct vmap_area, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) * During the traversal we also do some sanity check.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) * Trigger the BUG() if there are sides(left/right)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) * or full overlaps.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) if (va->va_start < tmp_va->va_end &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) va->va_end <= tmp_va->va_start)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) link = &(*link)->rb_left;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) else if (va->va_end > tmp_va->va_start &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) va->va_start >= tmp_va->va_end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) link = &(*link)->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) WARN(1, "vmalloc bug: 0x%lx-0x%lx overlaps with 0x%lx-0x%lx\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) va->va_start, va->va_end, tmp_va->va_start, tmp_va->va_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) } while (*link);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) *parent = &tmp_va->rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) return link;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) static __always_inline struct list_head *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) get_va_next_sibling(struct rb_node *parent, struct rb_node **link)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) struct list_head *list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) if (unlikely(!parent))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) * The red-black tree where we try to find VA neighbors
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) * before merging or inserting is empty, i.e. it means
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) * there is no free vmap space. Normally it does not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) * happen but we handle this case anyway.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) list = &rb_entry(parent, struct vmap_area, rb_node)->list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) return (&parent->rb_right == link ? list->next : list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) static __always_inline void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) link_va(struct vmap_area *va, struct rb_root *root,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) struct rb_node *parent, struct rb_node **link, struct list_head *head)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) * VA is still not in the list, but we can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596) * identify its future previous list_head node.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) if (likely(parent)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) head = &rb_entry(parent, struct vmap_area, rb_node)->list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) if (&parent->rb_right != link)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) head = head->prev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) /* Insert to the rb-tree */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) rb_link_node(&va->rb_node, parent, link);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) if (root == &free_vmap_area_root) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) * Some explanation here. Just perform simple insertion
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) * to the tree. We do not set va->subtree_max_size to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) * its current size before calling rb_insert_augmented().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) * It is because of we populate the tree from the bottom
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) * to parent levels when the node _is_ in the tree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) * Therefore we set subtree_max_size to zero after insertion,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) * to let __augment_tree_propagate_from() puts everything to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) * the correct order later on.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) rb_insert_augmented(&va->rb_node,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) root, &free_vmap_area_rb_augment_cb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) va->subtree_max_size = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) rb_insert_color(&va->rb_node, root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) /* Address-sort this list */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) list_add(&va->list, head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) static __always_inline void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) unlink_va(struct vmap_area *va, struct rb_root *root)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) if (WARN_ON(RB_EMPTY_NODE(&va->rb_node)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) if (root == &free_vmap_area_root)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) rb_erase_augmented(&va->rb_node,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) root, &free_vmap_area_rb_augment_cb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) rb_erase(&va->rb_node, root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) list_del(&va->list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) RB_CLEAR_NODE(&va->rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) #if DEBUG_AUGMENT_PROPAGATE_CHECK
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) augment_tree_propagate_check(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) unsigned long computed_size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652) list_for_each_entry(va, &free_vmap_area_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) computed_size = compute_subtree_max_size(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) if (computed_size != va->subtree_max_size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655) pr_emerg("tree is corrupted: %lu, %lu\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) va_size(va), va->subtree_max_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) * This function populates subtree_max_size from bottom to upper
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) * levels starting from VA point. The propagation must be done
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) * when VA size is modified by changing its va_start/va_end. Or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) * in case of newly inserting of VA to the tree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) * It means that __augment_tree_propagate_from() must be called:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668) * - After VA has been inserted to the tree(free path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) * - After VA has been shrunk(allocation path);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) * - After VA has been increased(merging path).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672) * Please note that, it does not mean that upper parent nodes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) * and their subtree_max_size are recalculated all the time up
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) * to the root node.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) * 4--8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) * /\
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) * / \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) * / \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) * 2--2 8--8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) * For example if we modify the node 4, shrinking it to 2, then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) * no any modification is required. If we shrink the node 2 to 1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) * its subtree_max_size is updated only, and set to 1. If we shrink
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) * the node 8 to 6, then its subtree_max_size is set to 6 and parent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) * node becomes 4--6.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) static __always_inline void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) augment_tree_propagate_from(struct vmap_area *va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) * Populate the tree from bottom towards the root until
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) * the calculated maximum available size of checked node
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) * is equal to its current one.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) free_vmap_area_rb_augment_cb_propagate(&va->rb_node, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698) #if DEBUG_AUGMENT_PROPAGATE_CHECK
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) augment_tree_propagate_check();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) insert_vmap_area(struct vmap_area *va,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) struct rb_root *root, struct list_head *head)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) struct rb_node **link;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) struct rb_node *parent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) link = find_va_links(va, root, NULL, &parent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) if (link)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) link_va(va, root, parent, link, head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) insert_vmap_area_augment(struct vmap_area *va,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) struct rb_node *from, struct rb_root *root,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718) struct list_head *head)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) struct rb_node **link;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) struct rb_node *parent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) if (from)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) link = find_va_links(va, NULL, from, &parent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) link = find_va_links(va, root, NULL, &parent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) if (link) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) link_va(va, root, parent, link, head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) augment_tree_propagate_from(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) * Merge de-allocated chunk of VA memory with previous
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) * and next free blocks. If coalesce is not done a new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) * free area is inserted. If VA has been merged, it is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738) * freed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) * Please note, it can return NULL in case of overlap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741) * ranges, followed by WARN() report. Despite it is a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) * buggy behaviour, a system can be alive and keep
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) * ongoing.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) static __always_inline struct vmap_area *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) merge_or_add_vmap_area(struct vmap_area *va,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) struct rb_root *root, struct list_head *head)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) struct vmap_area *sibling;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750) struct list_head *next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) struct rb_node **link;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) struct rb_node *parent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) bool merged = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) * Find a place in the tree where VA potentially will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) * inserted, unless it is merged with its sibling/siblings.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) link = find_va_links(va, root, NULL, &parent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) if (!link)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) * Get next node of VA to check if merging can be done.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) next = get_va_next_sibling(parent, link);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) if (unlikely(next == NULL))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) goto insert;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771) * start end
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) * | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) * |<------VA------>|<-----Next----->|
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) * | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) * start end
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) if (next != head) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778) sibling = list_entry(next, struct vmap_area, list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) if (sibling->va_start == va->va_end) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) sibling->va_start = va->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) /* Free vmap_area object. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) kmem_cache_free(vmap_area_cachep, va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) /* Point to the new merged area. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) va = sibling;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) merged = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792) * start end
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) * | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) * |<-----Prev----->|<------VA------>|
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) * | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) * start end
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) if (next->prev != head) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) sibling = list_entry(next->prev, struct vmap_area, list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) if (sibling->va_end == va->va_start) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) * If both neighbors are coalesced, it is important
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) * to unlink the "next" node first, followed by merging
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) * with "previous" one. Otherwise the tree might not be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) * fully populated if a sibling's augmented value is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) * "normalized" because of rotation operations.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) if (merged)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) unlink_va(va, root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) sibling->va_end = va->va_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) /* Free vmap_area object. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) kmem_cache_free(vmap_area_cachep, va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) /* Point to the new merged area. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) va = sibling;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) merged = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) insert:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) if (!merged)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) link_va(va, root, parent, link, head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) * Last step is to check and update the tree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) augment_tree_propagate_from(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) return va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) static __always_inline bool
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) is_within_this_va(struct vmap_area *va, unsigned long size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) unsigned long align, unsigned long vstart)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) unsigned long nva_start_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) if (va->va_start > vstart)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) nva_start_addr = ALIGN(va->va_start, align);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) nva_start_addr = ALIGN(vstart, align);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) /* Can be overflowed due to big size or alignment. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) if (nva_start_addr + size < nva_start_addr ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) nva_start_addr < vstart)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) return (nva_start_addr + size <= va->va_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) * Find the first free block(lowest start address) in the tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) * that will accomplish the request corresponding to passing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) * parameters.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) static __always_inline struct vmap_area *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) find_vmap_lowest_match(unsigned long size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) unsigned long align, unsigned long vstart)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) struct rb_node *node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) unsigned long length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) /* Start from the root. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) node = free_vmap_area_root.rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) /* Adjust the search size for alignment overhead. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) length = size + align - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) while (node) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) va = rb_entry(node, struct vmap_area, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) if (get_subtree_max_size(node->rb_left) >= length &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875) vstart < va->va_start) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876) node = node->rb_left;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) if (is_within_this_va(va, size, align, vstart))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879) return va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) * Does not make sense to go deeper towards the right
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) * sub-tree if it does not have a free block that is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884) * equal or bigger to the requested search length.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886) if (get_subtree_max_size(node->rb_right) >= length) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887) node = node->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) * OK. We roll back and find the first right sub-tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893) * that will satisfy the search criteria. It can happen
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894) * only once due to "vstart" restriction.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) while ((node = rb_parent(node))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897) va = rb_entry(node, struct vmap_area, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898) if (is_within_this_va(va, size, align, vstart))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899) return va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901) if (get_subtree_max_size(node->rb_right) >= length &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902) vstart <= va->va_start) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) node = node->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 912)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 913) #if DEBUG_AUGMENT_LOWEST_MATCH_CHECK
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 914) #include <linux/random.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916) static struct vmap_area *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917) find_vmap_lowest_linear_match(unsigned long size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918) unsigned long align, unsigned long vstart)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922) list_for_each_entry(va, &free_vmap_area_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923) if (!is_within_this_va(va, size, align, vstart))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926) return va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933) find_vmap_lowest_match_check(unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) struct vmap_area *va_1, *va_2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) unsigned long vstart;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937) unsigned int rnd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939) get_random_bytes(&rnd, sizeof(rnd));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) vstart = VMALLOC_START + rnd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 942) va_1 = find_vmap_lowest_match(size, 1, vstart);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 943) va_2 = find_vmap_lowest_linear_match(size, 1, vstart);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 944)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945) if (va_1 != va_2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) pr_emerg("not lowest: t: 0x%p, l: 0x%p, v: 0x%lx\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) va_1, va_2, vstart);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951) enum fit_type {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952) NOTHING_FIT = 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953) FL_FIT_TYPE = 1, /* full fit */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954) LE_FIT_TYPE = 2, /* left edge fit */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955) RE_FIT_TYPE = 3, /* right edge fit */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) NE_FIT_TYPE = 4 /* no edge fit */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 957) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 958)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 959) static __always_inline enum fit_type
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 960) classify_va_fit_type(struct vmap_area *va,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 961) unsigned long nva_start_addr, unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 962) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 963) enum fit_type type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 964)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 965) /* Check if it is within VA. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 966) if (nva_start_addr < va->va_start ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 967) nva_start_addr + size > va->va_end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 968) return NOTHING_FIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 969)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 970) /* Now classify. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 971) if (va->va_start == nva_start_addr) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 972) if (va->va_end == nva_start_addr + size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 973) type = FL_FIT_TYPE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 974) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 975) type = LE_FIT_TYPE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 976) } else if (va->va_end == nva_start_addr + size) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 977) type = RE_FIT_TYPE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 978) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 979) type = NE_FIT_TYPE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 980) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 981)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 982) return type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 983) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 984)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 985) static __always_inline int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 986) adjust_va_to_fit_type(struct vmap_area *va,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 987) unsigned long nva_start_addr, unsigned long size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 988) enum fit_type type)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 989) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 990) struct vmap_area *lva = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 991)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 992) if (type == FL_FIT_TYPE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 993) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 994) * No need to split VA, it fully fits.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 995) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 996) * | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 997) * V NVA V
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 998) * |---------------|
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 999) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1000) unlink_va(va, &free_vmap_area_root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1001) kmem_cache_free(vmap_area_cachep, va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1002) } else if (type == LE_FIT_TYPE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1003) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1004) * Split left edge of fit VA.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1005) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1006) * | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1007) * V NVA V R
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1008) * |-------|-------|
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1009) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1010) va->va_start += size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1011) } else if (type == RE_FIT_TYPE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1012) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1013) * Split right edge of fit VA.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1014) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1015) * | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1016) * L V NVA V
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1017) * |-------|-------|
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1018) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1019) va->va_end = nva_start_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1020) } else if (type == NE_FIT_TYPE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1021) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1022) * Split no edge of fit VA.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1023) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1024) * | |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1025) * L V NVA V R
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1026) * |---|-------|---|
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1027) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1028) lva = __this_cpu_xchg(ne_fit_preload_node, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1029) if (unlikely(!lva)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1030) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1031) * For percpu allocator we do not do any pre-allocation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1032) * and leave it as it is. The reason is it most likely
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1033) * never ends up with NE_FIT_TYPE splitting. In case of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1034) * percpu allocations offsets and sizes are aligned to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1035) * fixed align request, i.e. RE_FIT_TYPE and FL_FIT_TYPE
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1036) * are its main fitting cases.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1037) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1038) * There are a few exceptions though, as an example it is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1039) * a first allocation (early boot up) when we have "one"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1040) * big free space that has to be split.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1041) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1042) * Also we can hit this path in case of regular "vmap"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1043) * allocations, if "this" current CPU was not preloaded.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1044) * See the comment in alloc_vmap_area() why. If so, then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1045) * GFP_NOWAIT is used instead to get an extra object for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1046) * split purpose. That is rare and most time does not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1047) * occur.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1048) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1049) * What happens if an allocation gets failed. Basically,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1050) * an "overflow" path is triggered to purge lazily freed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1051) * areas to free some memory, then, the "retry" path is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1052) * triggered to repeat one more time. See more details
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1053) * in alloc_vmap_area() function.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1054) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1055) lva = kmem_cache_alloc(vmap_area_cachep, GFP_NOWAIT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1056) if (!lva)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1057) return -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1058) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1059)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1060) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1061) * Build the remainder.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1062) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1063) lva->va_start = va->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1064) lva->va_end = nva_start_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1065)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1066) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1067) * Shrink this VA to remaining size.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1068) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1069) va->va_start = nva_start_addr + size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1070) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1071) return -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1072) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1073)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1074) if (type != FL_FIT_TYPE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1075) augment_tree_propagate_from(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1076)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1077) if (lva) /* type == NE_FIT_TYPE */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1078) insert_vmap_area_augment(lva, &va->rb_node,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1079) &free_vmap_area_root, &free_vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1080) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1081)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1082) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1083) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1084)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1085) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1086) * Returns a start address of the newly allocated area, if success.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1087) * Otherwise a vend is returned that indicates failure.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1088) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1089) static __always_inline unsigned long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1090) __alloc_vmap_area(unsigned long size, unsigned long align,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1091) unsigned long vstart, unsigned long vend)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1092) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1093) unsigned long nva_start_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1094) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1095) enum fit_type type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1096) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1097)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1098) va = find_vmap_lowest_match(size, align, vstart);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1099) if (unlikely(!va))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1100) return vend;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1101)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1102) if (va->va_start > vstart)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1103) nva_start_addr = ALIGN(va->va_start, align);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1104) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1105) nva_start_addr = ALIGN(vstart, align);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1106)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1107) /* Check the "vend" restriction. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1108) if (nva_start_addr + size > vend)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1109) return vend;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1110)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1111) /* Classify what we have found. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1112) type = classify_va_fit_type(va, nva_start_addr, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1113) if (WARN_ON_ONCE(type == NOTHING_FIT))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1114) return vend;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1115)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1116) /* Update the free vmap_area. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1117) ret = adjust_va_to_fit_type(va, nva_start_addr, size, type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1118) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1119) return vend;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1120)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1121) #if DEBUG_AUGMENT_LOWEST_MATCH_CHECK
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1122) find_vmap_lowest_match_check(size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1123) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1124)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1125) return nva_start_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1126) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1127)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1128) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1129) * Free a region of KVA allocated by alloc_vmap_area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1130) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1131) static void free_vmap_area(struct vmap_area *va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1132) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1133) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1134) * Remove from the busy tree/list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1135) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1136) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1137) unlink_va(va, &vmap_area_root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1138) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1139)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1140) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1141) * Insert/Merge it back to the free tree/list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1142) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1143) spin_lock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1144) merge_or_add_vmap_area(va, &free_vmap_area_root, &free_vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1145) spin_unlock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1146) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1147)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1148) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1149) * Allocate a region of KVA of the specified size and alignment, within the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1150) * vstart and vend.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1151) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1152) static struct vmap_area *alloc_vmap_area(unsigned long size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1153) unsigned long align,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1154) unsigned long vstart, unsigned long vend,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1155) int node, gfp_t gfp_mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1156) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1157) struct vmap_area *va, *pva;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1158) unsigned long addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1159) int purged = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1160) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1161)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1162) BUG_ON(!size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1163) BUG_ON(offset_in_page(size));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1164) BUG_ON(!is_power_of_2(align));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1165)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1166) if (unlikely(!vmap_initialized))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1167) return ERR_PTR(-EBUSY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1169) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1170) gfp_mask = gfp_mask & GFP_RECLAIM_MASK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1171)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1172) va = kmem_cache_alloc_node(vmap_area_cachep, gfp_mask, node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1173) if (unlikely(!va))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1174) return ERR_PTR(-ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1175)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1176) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1177) * Only scan the relevant parts containing pointers to other objects
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1178) * to avoid false negatives.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1179) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1180) kmemleak_scan_area(&va->rb_node, SIZE_MAX, gfp_mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1182) retry:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1183) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1184) * Preload this CPU with one extra vmap_area object. It is used
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1185) * when fit type of free area is NE_FIT_TYPE. Please note, it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1186) * does not guarantee that an allocation occurs on a CPU that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1187) * is preloaded, instead we minimize the case when it is not.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1188) * It can happen because of cpu migration, because there is a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1189) * race until the below spinlock is taken.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1190) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1191) * The preload is done in non-atomic context, thus it allows us
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1192) * to use more permissive allocation masks to be more stable under
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1193) * low memory condition and high memory pressure. In rare case,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1194) * if not preloaded, GFP_NOWAIT is used.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1195) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1196) * Set "pva" to NULL here, because of "retry" path.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1197) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1198) pva = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1199)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1200) if (!this_cpu_read(ne_fit_preload_node))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1201) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1202) * Even if it fails we do not really care about that.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1203) * Just proceed as it is. If needed "overflow" path
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1204) * will refill the cache we allocate from.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1205) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1206) pva = kmem_cache_alloc_node(vmap_area_cachep, gfp_mask, node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1208) spin_lock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1209)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1210) if (pva && __this_cpu_cmpxchg(ne_fit_preload_node, NULL, pva))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1211) kmem_cache_free(vmap_area_cachep, pva);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1212)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1213) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1214) * If an allocation fails, the "vend" address is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1215) * returned. Therefore trigger the overflow path.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1216) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1217) addr = __alloc_vmap_area(size, align, vstart, vend);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1218) spin_unlock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1220) if (unlikely(addr == vend))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1221) goto overflow;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1222)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1223) va->va_start = addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1224) va->va_end = addr + size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1225) va->vm = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1226)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1227)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1228) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1229) insert_vmap_area(va, &vmap_area_root, &vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1230) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1231)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1232) BUG_ON(!IS_ALIGNED(va->va_start, align));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1233) BUG_ON(va->va_start < vstart);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1234) BUG_ON(va->va_end > vend);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1235)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1236) ret = kasan_populate_vmalloc(addr, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1237) if (ret) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1238) free_vmap_area(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1239) return ERR_PTR(ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1240) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1241)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1242) return va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1243)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1244) overflow:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1245) if (!purged) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1246) purge_vmap_area_lazy();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1247) purged = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1248) goto retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1249) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1250)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1251) if (gfpflags_allow_blocking(gfp_mask)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1252) unsigned long freed = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1253) blocking_notifier_call_chain(&vmap_notify_list, 0, &freed);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1254) if (freed > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1255) purged = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1256) goto retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1257) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1258) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1259)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1260) if (!(gfp_mask & __GFP_NOWARN) && printk_ratelimit())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1261) pr_warn("vmap allocation for size %lu failed: use vmalloc=<size> to increase size\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1262) size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1264) kmem_cache_free(vmap_area_cachep, va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1265) return ERR_PTR(-EBUSY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1266) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1267)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1268) int register_vmap_purge_notifier(struct notifier_block *nb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1269) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1270) return blocking_notifier_chain_register(&vmap_notify_list, nb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1271) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1272) EXPORT_SYMBOL_GPL(register_vmap_purge_notifier);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1273)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1274) int unregister_vmap_purge_notifier(struct notifier_block *nb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1275) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1276) return blocking_notifier_chain_unregister(&vmap_notify_list, nb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1277) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1278) EXPORT_SYMBOL_GPL(unregister_vmap_purge_notifier);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1279)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1280) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1281) * lazy_max_pages is the maximum amount of virtual address space we gather up
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1282) * before attempting to purge with a TLB flush.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1283) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1284) * There is a tradeoff here: a larger number will cover more kernel page tables
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1285) * and take slightly longer to purge, but it will linearly reduce the number of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1286) * global TLB flushes that must be performed. It would seem natural to scale
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1287) * this number up linearly with the number of CPUs (because vmapping activity
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1288) * could also scale linearly with the number of CPUs), however it is likely
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1289) * that in practice, workloads might be constrained in other ways that mean
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1290) * vmap activity will not scale linearly with CPUs. Also, I want to be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1291) * conservative and not introduce a big latency on huge systems, so go with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1292) * a less aggressive log scale. It will still be an improvement over the old
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1293) * code, and it will be simple to change the scale factor if we find that it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1294) * becomes a problem on bigger systems.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1295) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1296) static unsigned long lazy_max_pages(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1297) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1298) unsigned int log;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1299)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1300) log = fls(num_online_cpus());
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1301)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1302) return log * (32UL * 1024 * 1024 / PAGE_SIZE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1303) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1304)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1305) static atomic_long_t vmap_lazy_nr = ATOMIC_LONG_INIT(0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1306)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1307) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1308) * Serialize vmap purging. There is no actual criticial section protected
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1309) * by this look, but we want to avoid concurrent calls for performance
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1310) * reasons and to make the pcpu_get_vm_areas more deterministic.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1311) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1312) static DEFINE_MUTEX(vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1313)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1314) /* for per-CPU blocks */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1315) static void purge_fragmented_blocks_allcpus(void);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1316)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1317) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1318) * called before a call to iounmap() if the caller wants vm_area_struct's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1319) * immediately freed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1320) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1321) void set_iounmap_nonlazy(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1322) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1323) atomic_long_set(&vmap_lazy_nr, lazy_max_pages()+1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1324) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1325)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1326) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1327) * Purges all lazily-freed vmap areas.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1328) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1329) static bool __purge_vmap_area_lazy(unsigned long start, unsigned long end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1330) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1331) unsigned long resched_threshold;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1332) struct llist_node *valist;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1333) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1334) struct vmap_area *n_va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1335)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1336) lockdep_assert_held(&vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1337)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1338) valist = llist_del_all(&vmap_purge_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1339) if (unlikely(valist == NULL))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1340) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1341)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1342) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1343) * TODO: to calculate a flush range without looping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1344) * The list can be up to lazy_max_pages() elements.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1345) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1346) llist_for_each_entry(va, valist, purge_list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1347) if (va->va_start < start)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1348) start = va->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1349) if (va->va_end > end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1350) end = va->va_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1351) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1352)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1353) flush_tlb_kernel_range(start, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1354) resched_threshold = lazy_max_pages() << 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1355)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1356) spin_lock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1357) llist_for_each_entry_safe(va, n_va, valist, purge_list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1358) unsigned long nr = (va->va_end - va->va_start) >> PAGE_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1359) unsigned long orig_start = va->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1360) unsigned long orig_end = va->va_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1361)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1362) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1363) * Finally insert or merge lazily-freed area. It is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1364) * detached and there is no need to "unlink" it from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1365) * anything.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1366) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1367) va = merge_or_add_vmap_area(va, &free_vmap_area_root,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1368) &free_vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1369)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1370) if (!va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1371) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1372)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1373) if (is_vmalloc_or_module_addr((void *)orig_start))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1374) kasan_release_vmalloc(orig_start, orig_end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1375) va->va_start, va->va_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1376)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1377) atomic_long_sub(nr, &vmap_lazy_nr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1378)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1379) if (atomic_long_read(&vmap_lazy_nr) < resched_threshold)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1380) cond_resched_lock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1381) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1382) spin_unlock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1383) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1384) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1385)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1386) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1387) * Kick off a purge of the outstanding lazy areas. Don't bother if somebody
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1388) * is already purging.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1389) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1390) static void try_purge_vmap_area_lazy(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1391) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1392) if (mutex_trylock(&vmap_purge_lock)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1393) __purge_vmap_area_lazy(ULONG_MAX, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1394) mutex_unlock(&vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1395) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1396) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1397)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1398) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1399) * Kick off a purge of the outstanding lazy areas.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1400) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1401) static void purge_vmap_area_lazy(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1402) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1403) mutex_lock(&vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1404) purge_fragmented_blocks_allcpus();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1405) __purge_vmap_area_lazy(ULONG_MAX, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1406) mutex_unlock(&vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1407) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1408)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1409) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1410) * Free a vmap area, caller ensuring that the area has been unmapped
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1411) * and flush_cache_vunmap had been called for the correct range
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1412) * previously.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1413) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1414) static void free_vmap_area_noflush(struct vmap_area *va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1415) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1416) unsigned long nr_lazy;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1417)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1418) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1419) unlink_va(va, &vmap_area_root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1420) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1421)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1422) nr_lazy = atomic_long_add_return((va->va_end - va->va_start) >>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1423) PAGE_SHIFT, &vmap_lazy_nr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1424)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1425) /* After this point, we may free va at any time */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1426) llist_add(&va->purge_list, &vmap_purge_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1427)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1428) if (unlikely(nr_lazy > lazy_max_pages()))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1429) try_purge_vmap_area_lazy();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1430) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1431)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1432) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1433) * Free and unmap a vmap area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1434) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1435) static void free_unmap_vmap_area(struct vmap_area *va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1436) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1437) flush_cache_vunmap(va->va_start, va->va_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1438) unmap_kernel_range_noflush(va->va_start, va->va_end - va->va_start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1439) if (debug_pagealloc_enabled_static())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1440) flush_tlb_kernel_range(va->va_start, va->va_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1441)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1442) free_vmap_area_noflush(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1443) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1444)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1445) static struct vmap_area *find_vmap_area(unsigned long addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1446) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1447) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1448)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1449) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1450) va = __find_vmap_area(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1451) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1453) return va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1454) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1455)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1456) /*** Per cpu kva allocator ***/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1457)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1458) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1459) * vmap space is limited especially on 32 bit architectures. Ensure there is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1460) * room for at least 16 percpu vmap blocks per CPU.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1461) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1462) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1463) * If we had a constant VMALLOC_START and VMALLOC_END, we'd like to be able
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1464) * to #define VMALLOC_SPACE (VMALLOC_END-VMALLOC_START). Guess
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1465) * instead (we just need a rough idea)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1466) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1467) #if BITS_PER_LONG == 32
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1468) #define VMALLOC_SPACE (128UL*1024*1024)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1469) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1470) #define VMALLOC_SPACE (128UL*1024*1024*1024)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1471) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1472)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1473) #define VMALLOC_PAGES (VMALLOC_SPACE / PAGE_SIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1474) #define VMAP_MAX_ALLOC BITS_PER_LONG /* 256K with 4K pages */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1475) #define VMAP_BBMAP_BITS_MAX 1024 /* 4MB with 4K pages */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1476) #define VMAP_BBMAP_BITS_MIN (VMAP_MAX_ALLOC*2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1477) #define VMAP_MIN(x, y) ((x) < (y) ? (x) : (y)) /* can't use min() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1478) #define VMAP_MAX(x, y) ((x) > (y) ? (x) : (y)) /* can't use max() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1479) #define VMAP_BBMAP_BITS \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1480) VMAP_MIN(VMAP_BBMAP_BITS_MAX, \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1481) VMAP_MAX(VMAP_BBMAP_BITS_MIN, \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1482) VMALLOC_PAGES / roundup_pow_of_two(NR_CPUS) / 16))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1483)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1484) #define VMAP_BLOCK_SIZE (VMAP_BBMAP_BITS * PAGE_SIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1485)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1486) struct vmap_block_queue {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1487) spinlock_t lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1488) struct list_head free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1489) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1490)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1491) struct vmap_block {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1492) spinlock_t lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1493) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1494) unsigned long free, dirty;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1495) unsigned long dirty_min, dirty_max; /*< dirty range */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1496) struct list_head free_list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1497) struct rcu_head rcu_head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1498) struct list_head purge;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1499) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1500)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1501) /* Queue of free and dirty vmap blocks, for allocation and flushing purposes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1502) static DEFINE_PER_CPU(struct vmap_block_queue, vmap_block_queue);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1503)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1504) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1505) * XArray of vmap blocks, indexed by address, to quickly find a vmap block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1506) * in the free path. Could get rid of this if we change the API to return a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1507) * "cookie" from alloc, to be passed to free. But no big deal yet.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1508) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1509) static DEFINE_XARRAY(vmap_blocks);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1510)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1511) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1512) * We should probably have a fallback mechanism to allocate virtual memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1513) * out of partially filled vmap blocks. However vmap block sizing should be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1514) * fairly reasonable according to the vmalloc size, so it shouldn't be a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1515) * big problem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1516) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1517)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1518) static unsigned long addr_to_vb_idx(unsigned long addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1519) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1520) addr -= VMALLOC_START & ~(VMAP_BLOCK_SIZE-1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1521) addr /= VMAP_BLOCK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1522) return addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1523) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1524)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1525) static void *vmap_block_vaddr(unsigned long va_start, unsigned long pages_off)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1526) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1527) unsigned long addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1528)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1529) addr = va_start + (pages_off << PAGE_SHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1530) BUG_ON(addr_to_vb_idx(addr) != addr_to_vb_idx(va_start));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1531) return (void *)addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1532) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1533)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1534) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1535) * new_vmap_block - allocates new vmap_block and occupies 2^order pages in this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1536) * block. Of course pages number can't exceed VMAP_BBMAP_BITS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1537) * @order: how many 2^order pages should be occupied in newly allocated block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1538) * @gfp_mask: flags for the page level allocator
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1539) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1540) * Return: virtual address in a newly allocated block or ERR_PTR(-errno)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1541) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1542) static void *new_vmap_block(unsigned int order, gfp_t gfp_mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1543) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1544) struct vmap_block_queue *vbq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1545) struct vmap_block *vb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1546) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1547) unsigned long vb_idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1548) int node, err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1549) void *vaddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1550)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1551) node = numa_node_id();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1552)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1553) vb = kmalloc_node(sizeof(struct vmap_block),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1554) gfp_mask & GFP_RECLAIM_MASK, node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1555) if (unlikely(!vb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1556) return ERR_PTR(-ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1557)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1558) va = alloc_vmap_area(VMAP_BLOCK_SIZE, VMAP_BLOCK_SIZE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1559) VMALLOC_START, VMALLOC_END,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1560) node, gfp_mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1561) if (IS_ERR(va)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1562) kfree(vb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1563) return ERR_CAST(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1564) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1565)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1566) vaddr = vmap_block_vaddr(va->va_start, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1567) spin_lock_init(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1568) vb->va = va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1569) /* At least something should be left free */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1570) BUG_ON(VMAP_BBMAP_BITS <= (1UL << order));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1571) vb->free = VMAP_BBMAP_BITS - (1UL << order);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1572) vb->dirty = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1573) vb->dirty_min = VMAP_BBMAP_BITS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1574) vb->dirty_max = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1575) INIT_LIST_HEAD(&vb->free_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1576)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1577) vb_idx = addr_to_vb_idx(va->va_start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1578) err = xa_insert(&vmap_blocks, vb_idx, vb, gfp_mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1579) if (err) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1580) kfree(vb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1581) free_vmap_area(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1582) return ERR_PTR(err);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1583) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1584)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1585) vbq = &get_cpu_var(vmap_block_queue);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1586) spin_lock(&vbq->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1587) list_add_tail_rcu(&vb->free_list, &vbq->free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1588) spin_unlock(&vbq->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1589) put_cpu_var(vmap_block_queue);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1590)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1591) return vaddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1592) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1593)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1594) static void free_vmap_block(struct vmap_block *vb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1595) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1596) struct vmap_block *tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1597)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1598) tmp = xa_erase(&vmap_blocks, addr_to_vb_idx(vb->va->va_start));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1599) BUG_ON(tmp != vb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1600)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1601) free_vmap_area_noflush(vb->va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1602) kfree_rcu(vb, rcu_head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1603) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1604)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1605) static void purge_fragmented_blocks(int cpu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1606) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1607) LIST_HEAD(purge);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1608) struct vmap_block *vb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1609) struct vmap_block *n_vb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1610) struct vmap_block_queue *vbq = &per_cpu(vmap_block_queue, cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1611)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1612) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1613) list_for_each_entry_rcu(vb, &vbq->free, free_list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1614)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1615) if (!(vb->free + vb->dirty == VMAP_BBMAP_BITS && vb->dirty != VMAP_BBMAP_BITS))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1616) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1617)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1618) spin_lock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1619) if (vb->free + vb->dirty == VMAP_BBMAP_BITS && vb->dirty != VMAP_BBMAP_BITS) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1620) vb->free = 0; /* prevent further allocs after releasing lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1621) vb->dirty = VMAP_BBMAP_BITS; /* prevent purging it again */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1622) vb->dirty_min = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1623) vb->dirty_max = VMAP_BBMAP_BITS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1624) spin_lock(&vbq->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1625) list_del_rcu(&vb->free_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1626) spin_unlock(&vbq->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1627) spin_unlock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1628) list_add_tail(&vb->purge, &purge);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1629) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1630) spin_unlock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1631) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1632) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1633)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1634) list_for_each_entry_safe(vb, n_vb, &purge, purge) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1635) list_del(&vb->purge);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1636) free_vmap_block(vb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1637) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1638) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1639)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1640) static void purge_fragmented_blocks_allcpus(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1641) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1642) int cpu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1643)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1644) for_each_possible_cpu(cpu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1645) purge_fragmented_blocks(cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1646) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1647)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1648) static void *vb_alloc(unsigned long size, gfp_t gfp_mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1649) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1650) struct vmap_block_queue *vbq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1651) struct vmap_block *vb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1652) void *vaddr = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1653) unsigned int order;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1654)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1655) BUG_ON(offset_in_page(size));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1656) BUG_ON(size > PAGE_SIZE*VMAP_MAX_ALLOC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1657) if (WARN_ON(size == 0)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1658) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1659) * Allocating 0 bytes isn't what caller wants since
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1660) * get_order(0) returns funny result. Just warn and terminate
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1661) * early.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1662) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1663) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1664) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1665) order = get_order(size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1666)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1667) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1668) vbq = &get_cpu_var(vmap_block_queue);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1669) list_for_each_entry_rcu(vb, &vbq->free, free_list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1670) unsigned long pages_off;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1671)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1672) spin_lock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1673) if (vb->free < (1UL << order)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1674) spin_unlock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1675) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1676) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1677)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1678) pages_off = VMAP_BBMAP_BITS - vb->free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1679) vaddr = vmap_block_vaddr(vb->va->va_start, pages_off);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1680) vb->free -= 1UL << order;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1681) if (vb->free == 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1682) spin_lock(&vbq->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1683) list_del_rcu(&vb->free_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1684) spin_unlock(&vbq->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1685) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1686)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1687) spin_unlock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1688) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1689) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1690)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1691) put_cpu_var(vmap_block_queue);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1692) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1693)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1694) /* Allocate new block if nothing was found */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1695) if (!vaddr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1696) vaddr = new_vmap_block(order, gfp_mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1697)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1698) return vaddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1699) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1700)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1701) static void vb_free(unsigned long addr, unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1702) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1703) unsigned long offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1704) unsigned int order;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1705) struct vmap_block *vb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1706)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1707) BUG_ON(offset_in_page(size));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1708) BUG_ON(size > PAGE_SIZE*VMAP_MAX_ALLOC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1709)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1710) flush_cache_vunmap(addr, addr + size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1711)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1712) order = get_order(size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1713) offset = (addr & (VMAP_BLOCK_SIZE - 1)) >> PAGE_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1714) vb = xa_load(&vmap_blocks, addr_to_vb_idx(addr));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1715)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1716) unmap_kernel_range_noflush(addr, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1717)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1718) if (debug_pagealloc_enabled_static())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1719) flush_tlb_kernel_range(addr, addr + size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1720)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1721) spin_lock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1722)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1723) /* Expand dirty range */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1724) vb->dirty_min = min(vb->dirty_min, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1725) vb->dirty_max = max(vb->dirty_max, offset + (1UL << order));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1726)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1727) vb->dirty += 1UL << order;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1728) if (vb->dirty == VMAP_BBMAP_BITS) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1729) BUG_ON(vb->free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1730) spin_unlock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1731) free_vmap_block(vb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1732) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1733) spin_unlock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1734) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1735)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1736) static void _vm_unmap_aliases(unsigned long start, unsigned long end, int flush)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1737) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1738) int cpu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1739)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1740) if (unlikely(!vmap_initialized))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1741) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1742)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1743) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1744)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1745) for_each_possible_cpu(cpu) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1746) struct vmap_block_queue *vbq = &per_cpu(vmap_block_queue, cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1747) struct vmap_block *vb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1748)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1749) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1750) list_for_each_entry_rcu(vb, &vbq->free, free_list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1751) spin_lock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1752) if (vb->dirty && vb->dirty != VMAP_BBMAP_BITS) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1753) unsigned long va_start = vb->va->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1754) unsigned long s, e;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1755)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1756) s = va_start + (vb->dirty_min << PAGE_SHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1757) e = va_start + (vb->dirty_max << PAGE_SHIFT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1758)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1759) start = min(s, start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1760) end = max(e, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1761)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1762) flush = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1763) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1764) spin_unlock(&vb->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1765) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1766) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1767) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1768)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1769) mutex_lock(&vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1770) purge_fragmented_blocks_allcpus();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1771) if (!__purge_vmap_area_lazy(start, end) && flush)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1772) flush_tlb_kernel_range(start, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1773) mutex_unlock(&vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1774) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1775)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1776) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1777) * vm_unmap_aliases - unmap outstanding lazy aliases in the vmap layer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1778) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1779) * The vmap/vmalloc layer lazily flushes kernel virtual mappings primarily
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1780) * to amortize TLB flushing overheads. What this means is that any page you
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1781) * have now, may, in a former life, have been mapped into kernel virtual
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1782) * address by the vmap layer and so there might be some CPUs with TLB entries
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1783) * still referencing that page (additional to the regular 1:1 kernel mapping).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1784) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1785) * vm_unmap_aliases flushes all such lazy mappings. After it returns, we can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1786) * be sure that none of the pages we have control over will have any aliases
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1787) * from the vmap layer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1788) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1789) void vm_unmap_aliases(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1790) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1791) unsigned long start = ULONG_MAX, end = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1792) int flush = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1793)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1794) _vm_unmap_aliases(start, end, flush);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1795) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1796) EXPORT_SYMBOL_GPL(vm_unmap_aliases);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1797)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1798) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1799) * vm_unmap_ram - unmap linear kernel address space set up by vm_map_ram
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1800) * @mem: the pointer returned by vm_map_ram
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1801) * @count: the count passed to that vm_map_ram call (cannot unmap partial)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1802) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1803) void vm_unmap_ram(const void *mem, unsigned int count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1804) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1805) unsigned long size = (unsigned long)count << PAGE_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1806) unsigned long addr = (unsigned long)mem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1807) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1808)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1809) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1810) BUG_ON(!addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1811) BUG_ON(addr < VMALLOC_START);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1812) BUG_ON(addr > VMALLOC_END);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1813) BUG_ON(!PAGE_ALIGNED(addr));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1814)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1815) kasan_poison_vmalloc(mem, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1816)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1817) if (likely(count <= VMAP_MAX_ALLOC)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1818) debug_check_no_locks_freed(mem, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1819) vb_free(addr, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1820) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1821) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1822)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1823) va = find_vmap_area(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1824) BUG_ON(!va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1825) debug_check_no_locks_freed((void *)va->va_start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1826) (va->va_end - va->va_start));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1827) free_unmap_vmap_area(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1828) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1829) EXPORT_SYMBOL(vm_unmap_ram);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1830)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1831) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1832) * vm_map_ram - map pages linearly into kernel virtual address (vmalloc space)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1833) * @pages: an array of pointers to the pages to be mapped
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1834) * @count: number of pages
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1835) * @node: prefer to allocate data structures on this node
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1836) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1837) * If you use this function for less than VMAP_MAX_ALLOC pages, it could be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1838) * faster than vmap so it's good. But if you mix long-life and short-life
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1839) * objects with vm_map_ram(), it could consume lots of address space through
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1840) * fragmentation (especially on a 32bit machine). You could see failures in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1841) * the end. Please use this function for short-lived objects.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1842) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1843) * Returns: a pointer to the address that has been mapped, or %NULL on failure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1844) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1845) void *vm_map_ram(struct page **pages, unsigned int count, int node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1846) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1847) unsigned long size = (unsigned long)count << PAGE_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1848) unsigned long addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1849) void *mem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1850)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1851) if (likely(count <= VMAP_MAX_ALLOC)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1852) mem = vb_alloc(size, GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1853) if (IS_ERR(mem))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1854) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1855) addr = (unsigned long)mem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1856) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1857) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1858) va = alloc_vmap_area(size, PAGE_SIZE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1859) VMALLOC_START, VMALLOC_END, node, GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1860) if (IS_ERR(va))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1861) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1862)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1863) addr = va->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1864) mem = (void *)addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1865) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1866)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1867) kasan_unpoison_vmalloc(mem, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1868)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1869) if (map_kernel_range(addr, size, PAGE_KERNEL, pages) < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1870) vm_unmap_ram(mem, count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1871) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1872) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1873) return mem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1874) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1875) EXPORT_SYMBOL(vm_map_ram);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1876)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1877) static struct vm_struct *vmlist __initdata;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1878)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1879) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1880) * vm_area_add_early - add vmap area early during boot
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1881) * @vm: vm_struct to add
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1882) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1883) * This function is used to add fixed kernel vm area to vmlist before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1884) * vmalloc_init() is called. @vm->addr, @vm->size, and @vm->flags
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1885) * should contain proper values and the other fields should be zero.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1886) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1887) * DO NOT USE THIS FUNCTION UNLESS YOU KNOW WHAT YOU'RE DOING.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1888) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1889) void __init vm_area_add_early(struct vm_struct *vm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1890) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1891) struct vm_struct *tmp, **p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1892)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1893) BUG_ON(vmap_initialized);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1894) for (p = &vmlist; (tmp = *p) != NULL; p = &tmp->next) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1895) if (tmp->addr >= vm->addr) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1896) BUG_ON(tmp->addr < vm->addr + vm->size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1897) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1898) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1899) BUG_ON(tmp->addr + tmp->size > vm->addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1900) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1901) vm->next = *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1902) *p = vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1903) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1904)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1905) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1906) * vm_area_register_early - register vmap area early during boot
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1907) * @vm: vm_struct to register
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1908) * @align: requested alignment
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1909) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1910) * This function is used to register kernel vm area before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1911) * vmalloc_init() is called. @vm->size and @vm->flags should contain
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1912) * proper values on entry and other fields should be zero. On return,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1913) * vm->addr contains the allocated address.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1914) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1915) * DO NOT USE THIS FUNCTION UNLESS YOU KNOW WHAT YOU'RE DOING.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1916) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1917) void __init vm_area_register_early(struct vm_struct *vm, size_t align)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1918) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1919) static size_t vm_init_off __initdata;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1920) unsigned long addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1921)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1922) addr = ALIGN(VMALLOC_START + vm_init_off, align);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1923) vm_init_off = PFN_ALIGN(addr + vm->size) - VMALLOC_START;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1924)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1925) vm->addr = (void *)addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1926)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1927) vm_area_add_early(vm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1928) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1929)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1930) static void vmap_init_free_space(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1931) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1932) unsigned long vmap_start = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1933) const unsigned long vmap_end = ULONG_MAX;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1934) struct vmap_area *busy, *free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1935)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1936) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1937) * B F B B B F
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1938) * -|-----|.....|-----|-----|-----|.....|-
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1939) * | The KVA space |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1940) * |<--------------------------------->|
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1941) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1942) list_for_each_entry(busy, &vmap_area_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1943) if (busy->va_start - vmap_start > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1944) free = kmem_cache_zalloc(vmap_area_cachep, GFP_NOWAIT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1945) if (!WARN_ON_ONCE(!free)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1946) free->va_start = vmap_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1947) free->va_end = busy->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1948)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1949) insert_vmap_area_augment(free, NULL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1950) &free_vmap_area_root,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1951) &free_vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1952) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1953) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1954)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1955) vmap_start = busy->va_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1956) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1957)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1958) if (vmap_end - vmap_start > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1959) free = kmem_cache_zalloc(vmap_area_cachep, GFP_NOWAIT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1960) if (!WARN_ON_ONCE(!free)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1961) free->va_start = vmap_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1962) free->va_end = vmap_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1963)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1964) insert_vmap_area_augment(free, NULL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1965) &free_vmap_area_root,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1966) &free_vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1967) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1968) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1969) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1970)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1971) void __init vmalloc_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1972) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1973) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1974) struct vm_struct *tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1975) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1976)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1977) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1978) * Create the cache for vmap_area objects.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1979) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1980) vmap_area_cachep = KMEM_CACHE(vmap_area, SLAB_PANIC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1981)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1982) for_each_possible_cpu(i) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1983) struct vmap_block_queue *vbq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1984) struct vfree_deferred *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1985)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1986) vbq = &per_cpu(vmap_block_queue, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1987) spin_lock_init(&vbq->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1988) INIT_LIST_HEAD(&vbq->free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1989) p = &per_cpu(vfree_deferred, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1990) init_llist_head(&p->list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1991) INIT_WORK(&p->wq, free_work);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1992) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1993)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1994) /* Import existing vmlist entries. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1995) for (tmp = vmlist; tmp; tmp = tmp->next) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1996) va = kmem_cache_zalloc(vmap_area_cachep, GFP_NOWAIT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1997) if (WARN_ON_ONCE(!va))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1998) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1999)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2000) va->va_start = (unsigned long)tmp->addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2001) va->va_end = va->va_start + tmp->size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2002) va->vm = tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2003) insert_vmap_area(va, &vmap_area_root, &vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2004) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2005)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2006) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2007) * Now we can initialize a free vmap space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2008) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2009) vmap_init_free_space();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2010) vmap_initialized = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2011) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2012)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2013) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2014) * unmap_kernel_range - unmap kernel VM area and flush cache and TLB
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2015) * @addr: start of the VM area to unmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2016) * @size: size of the VM area to unmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2017) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2018) * Similar to unmap_kernel_range_noflush() but flushes vcache before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2019) * the unmapping and tlb after.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2020) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2021) void unmap_kernel_range(unsigned long addr, unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2022) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2023) unsigned long end = addr + size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2024)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2025) flush_cache_vunmap(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2026) unmap_kernel_range_noflush(addr, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2027) flush_tlb_kernel_range(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2028) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2029)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2030) static inline void setup_vmalloc_vm_locked(struct vm_struct *vm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2031) struct vmap_area *va, unsigned long flags, const void *caller)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2032) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2033) vm->flags = flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2034) vm->addr = (void *)va->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2035) vm->size = va->va_end - va->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2036) vm->caller = caller;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2037) va->vm = vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2038) trace_android_vh_save_vmalloc_stack(flags, vm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2039) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2040)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2041) static void setup_vmalloc_vm(struct vm_struct *vm, struct vmap_area *va,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2042) unsigned long flags, const void *caller)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2043) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2044) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2045) setup_vmalloc_vm_locked(vm, va, flags, caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2046) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2047) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2048)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2049) static void clear_vm_uninitialized_flag(struct vm_struct *vm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2050) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2051) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2052) * Before removing VM_UNINITIALIZED,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2053) * we should make sure that vm has proper values.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2054) * Pair with smp_rmb() in show_numa_info().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2055) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2056) smp_wmb();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2057) vm->flags &= ~VM_UNINITIALIZED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2058) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2059)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2060) static struct vm_struct *__get_vm_area_node(unsigned long size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2061) unsigned long align, unsigned long flags, unsigned long start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2062) unsigned long end, int node, gfp_t gfp_mask, const void *caller)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2063) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2064) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2065) struct vm_struct *area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2066) unsigned long requested_size = size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2067)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2068) BUG_ON(in_interrupt());
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2069) size = PAGE_ALIGN(size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2070) if (unlikely(!size))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2071) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2072)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2073) if (flags & VM_IOREMAP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2074) align = 1ul << clamp_t(int, get_count_order_long(size),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2075) PAGE_SHIFT, IOREMAP_MAX_ORDER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2076)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2077) area = kzalloc_node(sizeof(*area), gfp_mask & GFP_RECLAIM_MASK, node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2078) if (unlikely(!area))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2079) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2080)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2081) if (!(flags & VM_NO_GUARD))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2082) size += PAGE_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2083)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2084) va = alloc_vmap_area(size, align, start, end, node, gfp_mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2085) if (IS_ERR(va)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2086) kfree(area);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2087) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2088) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2089)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2090) kasan_unpoison_vmalloc((void *)va->va_start, requested_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2091)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2092) setup_vmalloc_vm(area, va, flags, caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2093)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2094) return area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2095) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2096)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2097) struct vm_struct *__get_vm_area_caller(unsigned long size, unsigned long flags,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2098) unsigned long start, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2099) const void *caller)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2100) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2101) return __get_vm_area_node(size, 1, flags, start, end, NUMA_NO_NODE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2102) GFP_KERNEL, caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2103) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2104) EXPORT_SYMBOL_GPL(__get_vm_area_caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2105)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2106) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2107) * get_vm_area - reserve a contiguous kernel virtual area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2108) * @size: size of the area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2109) * @flags: %VM_IOREMAP for I/O mappings or VM_ALLOC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2110) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2111) * Search an area of @size in the kernel virtual mapping area,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2112) * and reserved it for out purposes. Returns the area descriptor
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2113) * on success or %NULL on failure.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2114) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2115) * Return: the area descriptor on success or %NULL on failure.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2116) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2117) struct vm_struct *get_vm_area(unsigned long size, unsigned long flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2118) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2119) return __get_vm_area_node(size, 1, flags, VMALLOC_START, VMALLOC_END,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2120) NUMA_NO_NODE, GFP_KERNEL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2121) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2122) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2123)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2124) struct vm_struct *get_vm_area_caller(unsigned long size, unsigned long flags,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2125) const void *caller)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2126) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2127) return __get_vm_area_node(size, 1, flags, VMALLOC_START, VMALLOC_END,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2128) NUMA_NO_NODE, GFP_KERNEL, caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2129) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2130)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2131) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2132) * find_vm_area - find a continuous kernel virtual area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2133) * @addr: base address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2134) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2135) * Search for the kernel VM area starting at @addr, and return it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2136) * It is up to the caller to do all required locking to keep the returned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2137) * pointer valid.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2138) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2139) * Return: the area descriptor on success or %NULL on failure.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2140) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2141) struct vm_struct *find_vm_area(const void *addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2142) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2143) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2144)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2145) va = find_vmap_area((unsigned long)addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2146) if (!va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2147) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2148)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2149) return va->vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2150) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2151)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2152) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2153) * remove_vm_area - find and remove a continuous kernel virtual area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2154) * @addr: base address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2155) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2156) * Search for the kernel VM area starting at @addr, and remove it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2157) * This function returns the found VM area, but using it is NOT safe
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2158) * on SMP machines, except for its size or flags.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2159) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2160) * Return: the area descriptor on success or %NULL on failure.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2161) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2162) struct vm_struct *remove_vm_area(const void *addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2163) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2164) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2165)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2166) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2167)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2168) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2169) va = __find_vmap_area((unsigned long)addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2170) if (va && va->vm) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2171) struct vm_struct *vm = va->vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2172)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2173) va->vm = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2174) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2175)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2176) kasan_free_shadow(vm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2177) free_unmap_vmap_area(va);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2178)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2179) return vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2180) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2182) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2183) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2184) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2185)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2186) static inline void set_area_direct_map(const struct vm_struct *area,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2187) int (*set_direct_map)(struct page *page))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2188) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2189) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2190)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2191) for (i = 0; i < area->nr_pages; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2192) if (page_address(area->pages[i]))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2193) set_direct_map(area->pages[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2194) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2195)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2196) /* Handle removing and resetting vm mappings related to the vm_struct. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2197) static void vm_remove_mappings(struct vm_struct *area, int deallocate_pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2198) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2199) unsigned long start = ULONG_MAX, end = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2200) int flush_reset = area->flags & VM_FLUSH_RESET_PERMS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2201) int flush_dmap = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2202) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2203)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2204) remove_vm_area(area->addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2205)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2206) /* If this is not VM_FLUSH_RESET_PERMS memory, no need for the below. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2207) if (!flush_reset)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2208) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2209)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2210) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2211) * If not deallocating pages, just do the flush of the VM area and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2212) * return.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2213) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2214) if (!deallocate_pages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2215) vm_unmap_aliases();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2216) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2217) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2218)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2219) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2220) * If execution gets here, flush the vm mapping and reset the direct
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2221) * map. Find the start and end range of the direct mappings to make sure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2222) * the vm_unmap_aliases() flush includes the direct map.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2223) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2224) for (i = 0; i < area->nr_pages; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2225) unsigned long addr = (unsigned long)page_address(area->pages[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2226) if (addr) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2227) start = min(addr, start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2228) end = max(addr + PAGE_SIZE, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2229) flush_dmap = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2230) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2231) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2232)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2233) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2234) * Set direct map to something invalid so that it won't be cached if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2235) * there are any accesses after the TLB flush, then flush the TLB and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2236) * reset the direct map permissions to the default.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2237) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2238) set_area_direct_map(area, set_direct_map_invalid_noflush);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2239) _vm_unmap_aliases(start, end, flush_dmap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2240) set_area_direct_map(area, set_direct_map_default_noflush);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2241) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2242)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2243) static void __vunmap(const void *addr, int deallocate_pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2244) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2245) struct vm_struct *area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2246)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2247) if (!addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2248) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2249)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2250) if (WARN(!PAGE_ALIGNED(addr), "Trying to vfree() bad address (%p)\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2251) addr))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2252) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2253)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2254) area = find_vm_area(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2255) if (unlikely(!area)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2256) WARN(1, KERN_ERR "Trying to vfree() nonexistent vm area (%p)\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2257) addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2258) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2259) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2260)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2261) debug_check_no_locks_freed(area->addr, get_vm_area_size(area));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2262) debug_check_no_obj_freed(area->addr, get_vm_area_size(area));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2264) kasan_poison_vmalloc(area->addr, get_vm_area_size(area));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2265)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2266) vm_remove_mappings(area, deallocate_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2267)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2268) if (deallocate_pages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2269) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2270)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2271) for (i = 0; i < area->nr_pages; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2272) struct page *page = area->pages[i];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2273)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2274) BUG_ON(!page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2275) __free_pages(page, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2276) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2277) atomic_long_sub(area->nr_pages, &nr_vmalloc_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2278)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2279) kvfree(area->pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2280) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2281)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2282) kfree(area);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2283) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2284) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2285)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2286) static inline void __vfree_deferred(const void *addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2287) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2288) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2289) * Use raw_cpu_ptr() because this can be called from preemptible
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2290) * context. Preemption is absolutely fine here, because the llist_add()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2291) * implementation is lockless, so it works even if we are adding to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2292) * another cpu's list. schedule_work() should be fine with this too.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2293) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2294) struct vfree_deferred *p = raw_cpu_ptr(&vfree_deferred);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2295)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2296) if (llist_add((struct llist_node *)addr, &p->list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2297) schedule_work(&p->wq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2298) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2299)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2300) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2301) * vfree_atomic - release memory allocated by vmalloc()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2302) * @addr: memory base address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2303) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2304) * This one is just like vfree() but can be called in any atomic context
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2305) * except NMIs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2306) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2307) void vfree_atomic(const void *addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2308) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2309) BUG_ON(in_nmi());
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2310)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2311) kmemleak_free(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2312)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2313) if (!addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2314) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2315) __vfree_deferred(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2316) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2317)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2318) static void __vfree(const void *addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2319) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2320) if (unlikely(in_interrupt()))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2321) __vfree_deferred(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2322) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2323) __vunmap(addr, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2324) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2325)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2326) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2327) * vfree - Release memory allocated by vmalloc()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2328) * @addr: Memory base address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2329) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2330) * Free the virtually continuous memory area starting at @addr, as obtained
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2331) * from one of the vmalloc() family of APIs. This will usually also free the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2332) * physical memory underlying the virtual allocation, but that memory is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2333) * reference counted, so it will not be freed until the last user goes away.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2334) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2335) * If @addr is NULL, no operation is performed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2336) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2337) * Context:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2338) * May sleep if called *not* from interrupt context.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2339) * Must not be called in NMI context (strictly speaking, it could be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2340) * if we have CONFIG_ARCH_HAVE_NMI_SAFE_CMPXCHG, but making the calling
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2341) * conventions for vfree() arch-depenedent would be a really bad idea).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2342) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2343) void vfree(const void *addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2344) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2345) BUG_ON(in_nmi());
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2346)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2347) kmemleak_free(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2348)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2349) might_sleep_if(!in_interrupt());
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2350)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2351) if (!addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2352) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2353)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2354) __vfree(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2355) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2356) EXPORT_SYMBOL(vfree);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2357)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2358) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2359) * vunmap - release virtual mapping obtained by vmap()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2360) * @addr: memory base address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2361) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2362) * Free the virtually contiguous memory area starting at @addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2363) * which was created from the page array passed to vmap().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2364) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2365) * Must not be called in interrupt context.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2366) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2367) void vunmap(const void *addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2368) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2369) BUG_ON(in_interrupt());
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2370) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2371) if (addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2372) __vunmap(addr, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2373) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2374) EXPORT_SYMBOL(vunmap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2375)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2376) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2377) * vmap - map an array of pages into virtually contiguous space
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2378) * @pages: array of page pointers
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2379) * @count: number of pages to map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2380) * @flags: vm_area->flags
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2381) * @prot: page protection for the mapping
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2382) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2383) * Maps @count pages from @pages into contiguous kernel virtual space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2384) * If @flags contains %VM_MAP_PUT_PAGES the ownership of the pages array itself
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2385) * (which must be kmalloc or vmalloc memory) and one reference per pages in it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2386) * are transferred from the caller to vmap(), and will be freed / dropped when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2387) * vfree() is called on the return value.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2388) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2389) * Return: the address of the area or %NULL on failure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2390) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2391) void *vmap(struct page **pages, unsigned int count,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2392) unsigned long flags, pgprot_t prot)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2393) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2394) struct vm_struct *area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2395) unsigned long size; /* In bytes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2396)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2397) might_sleep();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2398)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2399) if (count > totalram_pages())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2400) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2401)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2402) size = (unsigned long)count << PAGE_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2403) area = get_vm_area_caller(size, flags, __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2404) if (!area)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2405) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2406)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2407) if (map_kernel_range((unsigned long)area->addr, size, pgprot_nx(prot),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2408) pages) < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2409) vunmap(area->addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2410) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2411) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2412)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2413) if (flags & VM_MAP_PUT_PAGES) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2414) area->pages = pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2415) area->nr_pages = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2416) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2417) return area->addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2418) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2419) EXPORT_SYMBOL(vmap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2420)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2421) #ifdef CONFIG_VMAP_PFN
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2422) struct vmap_pfn_data {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2423) unsigned long *pfns;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2424) pgprot_t prot;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2425) unsigned int idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2426) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2427)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2428) static int vmap_pfn_apply(pte_t *pte, unsigned long addr, void *private)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2429) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2430) struct vmap_pfn_data *data = private;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2431)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2432) if (WARN_ON_ONCE(pfn_valid(data->pfns[data->idx])))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2433) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2434) *pte = pte_mkspecial(pfn_pte(data->pfns[data->idx++], data->prot));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2435) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2436) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2437)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2438) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2439) * vmap_pfn - map an array of PFNs into virtually contiguous space
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2440) * @pfns: array of PFNs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2441) * @count: number of pages to map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2442) * @prot: page protection for the mapping
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2443) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2444) * Maps @count PFNs from @pfns into contiguous kernel virtual space and returns
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2445) * the start address of the mapping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2446) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2447) void *vmap_pfn(unsigned long *pfns, unsigned int count, pgprot_t prot)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2448) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2449) struct vmap_pfn_data data = { .pfns = pfns, .prot = pgprot_nx(prot) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2450) struct vm_struct *area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2451)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2452) area = get_vm_area_caller(count * PAGE_SIZE, VM_IOREMAP,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2453) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2454) if (!area)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2455) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2456) if (apply_to_page_range(&init_mm, (unsigned long)area->addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2457) count * PAGE_SIZE, vmap_pfn_apply, &data)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2458) free_vm_area(area);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2459) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2460) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2461) return area->addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2462) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2463) EXPORT_SYMBOL_GPL(vmap_pfn);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2464) #endif /* CONFIG_VMAP_PFN */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2465)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2466) static void *__vmalloc_area_node(struct vm_struct *area, gfp_t gfp_mask,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2467) pgprot_t prot, int node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2468) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2469) const gfp_t nested_gfp = (gfp_mask & GFP_RECLAIM_MASK) | __GFP_ZERO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2470) unsigned int nr_pages = get_vm_area_size(area) >> PAGE_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2471) unsigned int array_size = nr_pages * sizeof(struct page *), i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2472) struct page **pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2473)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2474) gfp_mask |= __GFP_NOWARN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2475) if (!(gfp_mask & (GFP_DMA | GFP_DMA32)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2476) gfp_mask |= __GFP_HIGHMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2477)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2478) /* Please note that the recursion is strictly bounded. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2479) if (array_size > PAGE_SIZE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2480) pages = __vmalloc_node(array_size, 1, nested_gfp, node,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2481) area->caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2482) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2483) pages = kmalloc_node(array_size, nested_gfp, node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2484) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2485)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2486) if (!pages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2487) remove_vm_area(area->addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2488) kfree(area);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2489) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2490) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2491)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2492) area->pages = pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2493) area->nr_pages = nr_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2494)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2495) for (i = 0; i < area->nr_pages; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2496) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2497)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2498) if (node == NUMA_NO_NODE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2499) page = alloc_page(gfp_mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2500) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2501) page = alloc_pages_node(node, gfp_mask, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2502)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2503) if (unlikely(!page)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2504) /* Successfully allocated i pages, free them in __vfree() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2505) area->nr_pages = i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2506) atomic_long_add(area->nr_pages, &nr_vmalloc_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2507) goto fail;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2508) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2509) area->pages[i] = page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2510) if (gfpflags_allow_blocking(gfp_mask))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2511) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2512) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2513) atomic_long_add(area->nr_pages, &nr_vmalloc_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2514)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2515) if (map_kernel_range((unsigned long)area->addr, get_vm_area_size(area),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2516) prot, pages) < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2517) goto fail;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2518)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2519) return area->addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2520)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2521) fail:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2522) warn_alloc(gfp_mask, NULL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2523) "vmalloc: allocation failure, allocated %ld of %ld bytes",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2524) (area->nr_pages*PAGE_SIZE), area->size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2525) __vfree(area->addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2526) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2527) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2528)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2529) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2530) * __vmalloc_node_range - allocate virtually contiguous memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2531) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2532) * @align: desired alignment
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2533) * @start: vm area range start
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2534) * @end: vm area range end
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2535) * @gfp_mask: flags for the page level allocator
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2536) * @prot: protection mask for the allocated pages
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2537) * @vm_flags: additional vm area flags (e.g. %VM_NO_GUARD)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2538) * @node: node to use for allocation or NUMA_NO_NODE
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2539) * @caller: caller's return address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2540) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2541) * Allocate enough pages to cover @size from the page level
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2542) * allocator with @gfp_mask flags. Map them into contiguous
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2543) * kernel virtual space, using a pagetable protection of @prot.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2544) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2545) * Return: the address of the area or %NULL on failure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2546) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2547) void *__vmalloc_node_range(unsigned long size, unsigned long align,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2548) unsigned long start, unsigned long end, gfp_t gfp_mask,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2549) pgprot_t prot, unsigned long vm_flags, int node,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2550) const void *caller)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2551) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2552) struct vm_struct *area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2553) void *addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2554) unsigned long real_size = size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2555)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2556) size = PAGE_ALIGN(size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2557) if (!size || (size >> PAGE_SHIFT) > totalram_pages())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2558) goto fail;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2559)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2560) area = __get_vm_area_node(real_size, align, VM_ALLOC | VM_UNINITIALIZED |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2561) vm_flags, start, end, node, gfp_mask, caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2562) if (!area)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2563) goto fail;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2564)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2565) addr = __vmalloc_area_node(area, gfp_mask, prot, node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2566) if (!addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2567) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2568)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2569) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2570) * In this function, newly allocated vm_struct has VM_UNINITIALIZED
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2571) * flag. It means that vm_struct is not fully initialized.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2572) * Now, it is fully initialized, so remove this flag here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2573) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2574) clear_vm_uninitialized_flag(area);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2575)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2576) kmemleak_vmalloc(area, size, gfp_mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2577)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2578) return addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2579)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2580) fail:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2581) warn_alloc(gfp_mask, NULL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2582) "vmalloc: allocation failure: %lu bytes", real_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2583) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2584) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2585)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2586) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2587) * __vmalloc_node - allocate virtually contiguous memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2588) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2589) * @align: desired alignment
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2590) * @gfp_mask: flags for the page level allocator
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2591) * @node: node to use for allocation or NUMA_NO_NODE
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2592) * @caller: caller's return address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2593) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2594) * Allocate enough pages to cover @size from the page level allocator with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2595) * @gfp_mask flags. Map them into contiguous kernel virtual space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2596) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2597) * Reclaim modifiers in @gfp_mask - __GFP_NORETRY, __GFP_RETRY_MAYFAIL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2598) * and __GFP_NOFAIL are not supported
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2599) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2600) * Any use of gfp flags outside of GFP_KERNEL should be consulted
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2601) * with mm people.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2602) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2603) * Return: pointer to the allocated memory or %NULL on error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2604) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2605) void *__vmalloc_node(unsigned long size, unsigned long align,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2606) gfp_t gfp_mask, int node, const void *caller)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2607) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2608) return __vmalloc_node_range(size, align, VMALLOC_START, VMALLOC_END,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2609) gfp_mask, PAGE_KERNEL, 0, node, caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2610) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2611) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2612) * This is only for performance analysis of vmalloc and stress purpose.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2613) * It is required by vmalloc test module, therefore do not use it other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2614) * than that.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2615) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2616) #ifdef CONFIG_TEST_VMALLOC_MODULE
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2617) EXPORT_SYMBOL_GPL(__vmalloc_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2618) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2619)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2620) void *__vmalloc(unsigned long size, gfp_t gfp_mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2621) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2622) return __vmalloc_node(size, 1, gfp_mask, NUMA_NO_NODE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2623) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2624) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2625) EXPORT_SYMBOL(__vmalloc);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2626)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2627) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2628) * vmalloc - allocate virtually contiguous memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2629) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2630) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2631) * Allocate enough pages to cover @size from the page level
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2632) * allocator and map them into contiguous kernel virtual space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2633) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2634) * For tight control over page level allocator and protection flags
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2635) * use __vmalloc() instead.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2636) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2637) * Return: pointer to the allocated memory or %NULL on error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2638) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2639) void *vmalloc(unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2640) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2641) return __vmalloc_node(size, 1, GFP_KERNEL, NUMA_NO_NODE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2642) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2643) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2644) EXPORT_SYMBOL(vmalloc);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2645)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2646) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2647) * vzalloc - allocate virtually contiguous memory with zero fill
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2648) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2649) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2650) * Allocate enough pages to cover @size from the page level
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2651) * allocator and map them into contiguous kernel virtual space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2652) * The memory allocated is set to zero.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2653) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2654) * For tight control over page level allocator and protection flags
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2655) * use __vmalloc() instead.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2656) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2657) * Return: pointer to the allocated memory or %NULL on error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2658) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2659) void *vzalloc(unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2660) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2661) return __vmalloc_node(size, 1, GFP_KERNEL | __GFP_ZERO, NUMA_NO_NODE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2662) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2663) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2664) EXPORT_SYMBOL(vzalloc);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2665)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2666) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2667) * vmalloc_user - allocate zeroed virtually contiguous memory for userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2668) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2669) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2670) * The resulting memory area is zeroed so it can be mapped to userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2671) * without leaking data.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2672) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2673) * Return: pointer to the allocated memory or %NULL on error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2674) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2675) void *vmalloc_user(unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2676) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2677) return __vmalloc_node_range(size, SHMLBA, VMALLOC_START, VMALLOC_END,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2678) GFP_KERNEL | __GFP_ZERO, PAGE_KERNEL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2679) VM_USERMAP, NUMA_NO_NODE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2680) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2681) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2682) EXPORT_SYMBOL(vmalloc_user);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2683)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2684) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2685) * vmalloc_node - allocate memory on a specific node
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2686) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2687) * @node: numa node
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2688) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2689) * Allocate enough pages to cover @size from the page level
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2690) * allocator and map them into contiguous kernel virtual space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2691) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2692) * For tight control over page level allocator and protection flags
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2693) * use __vmalloc() instead.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2694) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2695) * Return: pointer to the allocated memory or %NULL on error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2696) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2697) void *vmalloc_node(unsigned long size, int node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2698) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2699) return __vmalloc_node(size, 1, GFP_KERNEL, node,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2700) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2701) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2702) EXPORT_SYMBOL(vmalloc_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2703)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2704) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2705) * vzalloc_node - allocate memory on a specific node with zero fill
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2706) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2707) * @node: numa node
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2708) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2709) * Allocate enough pages to cover @size from the page level
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2710) * allocator and map them into contiguous kernel virtual space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2711) * The memory allocated is set to zero.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2712) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2713) * Return: pointer to the allocated memory or %NULL on error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2714) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2715) void *vzalloc_node(unsigned long size, int node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2716) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2717) return __vmalloc_node(size, 1, GFP_KERNEL | __GFP_ZERO, node,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2718) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2719) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2720) EXPORT_SYMBOL(vzalloc_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2721)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2722) #if defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA32)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2723) #define GFP_VMALLOC32 (GFP_DMA32 | GFP_KERNEL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2724) #elif defined(CONFIG_64BIT) && defined(CONFIG_ZONE_DMA)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2725) #define GFP_VMALLOC32 (GFP_DMA | GFP_KERNEL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2726) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2727) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2728) * 64b systems should always have either DMA or DMA32 zones. For others
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2729) * GFP_DMA32 should do the right thing and use the normal zone.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2730) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2731) #define GFP_VMALLOC32 GFP_DMA32 | GFP_KERNEL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2732) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2733)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2734) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2735) * vmalloc_32 - allocate virtually contiguous memory (32bit addressable)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2736) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2737) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2738) * Allocate enough 32bit PA addressable pages to cover @size from the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2739) * page level allocator and map them into contiguous kernel virtual space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2740) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2741) * Return: pointer to the allocated memory or %NULL on error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2742) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2743) void *vmalloc_32(unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2744) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2745) return __vmalloc_node(size, 1, GFP_VMALLOC32, NUMA_NO_NODE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2746) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2747) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2748) EXPORT_SYMBOL(vmalloc_32);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2749)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2750) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2751) * vmalloc_32_user - allocate zeroed virtually contiguous 32bit memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2752) * @size: allocation size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2753) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2754) * The resulting memory area is 32bit addressable and zeroed so it can be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2755) * mapped to userspace without leaking data.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2756) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2757) * Return: pointer to the allocated memory or %NULL on error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2758) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2759) void *vmalloc_32_user(unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2760) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2761) return __vmalloc_node_range(size, SHMLBA, VMALLOC_START, VMALLOC_END,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2762) GFP_VMALLOC32 | __GFP_ZERO, PAGE_KERNEL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2763) VM_USERMAP, NUMA_NO_NODE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2764) __builtin_return_address(0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2765) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2766) EXPORT_SYMBOL(vmalloc_32_user);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2767)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2768) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2769) * small helper routine , copy contents to buf from addr.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2770) * If the page is not present, fill zero.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2771) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2772)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2773) static int aligned_vread(char *buf, char *addr, unsigned long count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2774) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2775) struct page *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2776) int copied = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2777)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2778) while (count) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2779) unsigned long offset, length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2780)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2781) offset = offset_in_page(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2782) length = PAGE_SIZE - offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2783) if (length > count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2784) length = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2785) p = vmalloc_to_page(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2786) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2787) * To do safe access to this _mapped_ area, we need
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2788) * lock. But adding lock here means that we need to add
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2789) * overhead of vmalloc()/vfree() calles for this _debug_
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2790) * interface, rarely used. Instead of that, we'll use
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2791) * kmap() and get small overhead in this access function.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2792) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2793) if (p) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2794) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2795) * we can expect USER0 is not used (see vread/vwrite's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2796) * function description)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2797) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2798) void *map = kmap_atomic(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2799) memcpy(buf, map + offset, length);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2800) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2801) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2802) memset(buf, 0, length);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2803)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2804) addr += length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2805) buf += length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2806) copied += length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2807) count -= length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2808) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2809) return copied;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2810) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2811)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2812) static int aligned_vwrite(char *buf, char *addr, unsigned long count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2813) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2814) struct page *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2815) int copied = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2816)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2817) while (count) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2818) unsigned long offset, length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2819)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2820) offset = offset_in_page(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2821) length = PAGE_SIZE - offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2822) if (length > count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2823) length = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2824) p = vmalloc_to_page(addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2825) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2826) * To do safe access to this _mapped_ area, we need
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2827) * lock. But adding lock here means that we need to add
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2828) * overhead of vmalloc()/vfree() calles for this _debug_
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2829) * interface, rarely used. Instead of that, we'll use
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2830) * kmap() and get small overhead in this access function.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2831) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2832) if (p) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2833) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2834) * we can expect USER0 is not used (see vread/vwrite's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2835) * function description)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2836) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2837) void *map = kmap_atomic(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2838) memcpy(map + offset, buf, length);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2839) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2840) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2841) addr += length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2842) buf += length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2843) copied += length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2844) count -= length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2845) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2846) return copied;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2847) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2848)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2849) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2850) * vread() - read vmalloc area in a safe way.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2851) * @buf: buffer for reading data
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2852) * @addr: vm address.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2853) * @count: number of bytes to be read.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2854) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2855) * This function checks that addr is a valid vmalloc'ed area, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2856) * copy data from that area to a given buffer. If the given memory range
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2857) * of [addr...addr+count) includes some valid address, data is copied to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2858) * proper area of @buf. If there are memory holes, they'll be zero-filled.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2859) * IOREMAP area is treated as memory hole and no copy is done.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2860) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2861) * If [addr...addr+count) doesn't includes any intersects with alive
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2862) * vm_struct area, returns 0. @buf should be kernel's buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2863) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2864) * Note: In usual ops, vread() is never necessary because the caller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2865) * should know vmalloc() area is valid and can use memcpy().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2866) * This is for routines which have to access vmalloc area without
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2867) * any information, as /dev/kmem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2868) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2869) * Return: number of bytes for which addr and buf should be increased
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2870) * (same number as @count) or %0 if [addr...addr+count) doesn't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2871) * include any intersection with valid vmalloc area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2872) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2873) long vread(char *buf, char *addr, unsigned long count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2874) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2875) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2876) struct vm_struct *vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2877) char *vaddr, *buf_start = buf;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2878) unsigned long buflen = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2879) unsigned long n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2880)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2881) /* Don't allow overflow */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2882) if ((unsigned long) addr + count < count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2883) count = -(unsigned long) addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2884)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2885) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2886) list_for_each_entry(va, &vmap_area_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2887) if (!count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2888) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2889)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2890) if (!va->vm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2891) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2892)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2893) vm = va->vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2894) vaddr = (char *) vm->addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2895) if (addr >= vaddr + get_vm_area_size(vm))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2896) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2897) while (addr < vaddr) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2898) if (count == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2899) goto finished;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2900) *buf = '\0';
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2901) buf++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2902) addr++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2903) count--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2904) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2905) n = vaddr + get_vm_area_size(vm) - addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2906) if (n > count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2907) n = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2908) if (!(vm->flags & VM_IOREMAP))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2909) aligned_vread(buf, addr, n);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2910) else /* IOREMAP area is treated as memory hole */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2911) memset(buf, 0, n);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2912) buf += n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2913) addr += n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2914) count -= n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2915) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2916) finished:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2917) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2918)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2919) if (buf == buf_start)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2920) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2921) /* zero-fill memory holes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2922) if (buf != buf_start + buflen)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2923) memset(buf, 0, buflen - (buf - buf_start));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2924)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2925) return buflen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2926) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2927)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2928) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2929) * vwrite() - write vmalloc area in a safe way.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2930) * @buf: buffer for source data
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2931) * @addr: vm address.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2932) * @count: number of bytes to be read.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2933) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2934) * This function checks that addr is a valid vmalloc'ed area, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2935) * copy data from a buffer to the given addr. If specified range of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2936) * [addr...addr+count) includes some valid address, data is copied from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2937) * proper area of @buf. If there are memory holes, no copy to hole.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2938) * IOREMAP area is treated as memory hole and no copy is done.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2939) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2940) * If [addr...addr+count) doesn't includes any intersects with alive
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2941) * vm_struct area, returns 0. @buf should be kernel's buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2942) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2943) * Note: In usual ops, vwrite() is never necessary because the caller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2944) * should know vmalloc() area is valid and can use memcpy().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2945) * This is for routines which have to access vmalloc area without
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2946) * any information, as /dev/kmem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2947) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2948) * Return: number of bytes for which addr and buf should be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2949) * increased (same number as @count) or %0 if [addr...addr+count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2950) * doesn't include any intersection with valid vmalloc area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2951) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2952) long vwrite(char *buf, char *addr, unsigned long count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2953) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2954) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2955) struct vm_struct *vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2956) char *vaddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2957) unsigned long n, buflen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2958) int copied = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2959)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2960) /* Don't allow overflow */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2961) if ((unsigned long) addr + count < count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2962) count = -(unsigned long) addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2963) buflen = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2964)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2965) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2966) list_for_each_entry(va, &vmap_area_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2967) if (!count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2968) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2969)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2970) if (!va->vm)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2971) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2972)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2973) vm = va->vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2974) vaddr = (char *) vm->addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2975) if (addr >= vaddr + get_vm_area_size(vm))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2976) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2977) while (addr < vaddr) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2978) if (count == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2979) goto finished;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2980) buf++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2981) addr++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2982) count--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2983) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2984) n = vaddr + get_vm_area_size(vm) - addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2985) if (n > count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2986) n = count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2987) if (!(vm->flags & VM_IOREMAP)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2988) aligned_vwrite(buf, addr, n);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2989) copied++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2990) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2991) buf += n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2992) addr += n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2993) count -= n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2994) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2995) finished:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2996) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2997) if (!copied)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2998) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2999) return buflen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3000) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3001)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3002) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3003) * remap_vmalloc_range_partial - map vmalloc pages to userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3004) * @vma: vma to cover
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3005) * @uaddr: target user address to start at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3006) * @kaddr: virtual address of vmalloc kernel memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3007) * @pgoff: offset from @kaddr to start at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3008) * @size: size of map area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3009) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3010) * Returns: 0 for success, -Exxx on failure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3011) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3012) * This function checks that @kaddr is a valid vmalloc'ed area,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3013) * and that it is big enough to cover the range starting at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3014) * @uaddr in @vma. Will return failure if that criteria isn't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3015) * met.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3016) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3017) * Similar to remap_pfn_range() (see mm/memory.c)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3018) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3019) int remap_vmalloc_range_partial(struct vm_area_struct *vma, unsigned long uaddr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3020) void *kaddr, unsigned long pgoff,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3021) unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3022) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3023) struct vm_struct *area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3024) unsigned long off;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3025) unsigned long end_index;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3026)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3027) if (check_shl_overflow(pgoff, PAGE_SHIFT, &off))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3028) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3029)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3030) size = PAGE_ALIGN(size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3031)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3032) if (!PAGE_ALIGNED(uaddr) || !PAGE_ALIGNED(kaddr))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3033) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3034)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3035) area = find_vm_area(kaddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3036) if (!area)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3037) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3038)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3039) if (!(area->flags & (VM_USERMAP | VM_DMA_COHERENT)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3040) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3041)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3042) if (check_add_overflow(size, off, &end_index) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3043) end_index > get_vm_area_size(area))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3044) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3045) kaddr += off;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3046)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3047) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3048) struct page *page = vmalloc_to_page(kaddr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3049) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3050)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3051) ret = vm_insert_page(vma, uaddr, page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3052) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3053) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3054)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3055) uaddr += PAGE_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3056) kaddr += PAGE_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3057) size -= PAGE_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3058) } while (size > 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3059)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3060) vma->vm_flags |= VM_DONTEXPAND | VM_DONTDUMP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3061)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3062) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3063) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3064) EXPORT_SYMBOL(remap_vmalloc_range_partial);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3065)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3066) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3067) * remap_vmalloc_range - map vmalloc pages to userspace
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3068) * @vma: vma to cover (map full range of vma)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3069) * @addr: vmalloc memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3070) * @pgoff: number of pages into addr before first page to map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3071) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3072) * Returns: 0 for success, -Exxx on failure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3073) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3074) * This function checks that addr is a valid vmalloc'ed area, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3075) * that it is big enough to cover the vma. Will return failure if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3076) * that criteria isn't met.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3077) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3078) * Similar to remap_pfn_range() (see mm/memory.c)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3079) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3080) int remap_vmalloc_range(struct vm_area_struct *vma, void *addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3081) unsigned long pgoff)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3082) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3083) return remap_vmalloc_range_partial(vma, vma->vm_start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3084) addr, pgoff,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3085) vma->vm_end - vma->vm_start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3086) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3087) EXPORT_SYMBOL(remap_vmalloc_range);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3088)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3089) void free_vm_area(struct vm_struct *area)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3090) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3091) struct vm_struct *ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3092) ret = remove_vm_area(area->addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3093) BUG_ON(ret != area);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3094) kfree(area);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3095) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3096) EXPORT_SYMBOL_GPL(free_vm_area);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3097)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3098) #ifdef CONFIG_SMP
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3099) static struct vmap_area *node_to_va(struct rb_node *n)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3100) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3101) return rb_entry_safe(n, struct vmap_area, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3102) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3103)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3104) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3105) * pvm_find_va_enclose_addr - find the vmap_area @addr belongs to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3106) * @addr: target address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3107) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3108) * Returns: vmap_area if it is found. If there is no such area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3109) * the first highest(reverse order) vmap_area is returned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3110) * i.e. va->va_start < addr && va->va_end < addr or NULL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3111) * if there are no any areas before @addr.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3112) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3113) static struct vmap_area *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3114) pvm_find_va_enclose_addr(unsigned long addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3115) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3116) struct vmap_area *va, *tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3117) struct rb_node *n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3118)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3119) n = free_vmap_area_root.rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3120) va = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3121)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3122) while (n) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3123) tmp = rb_entry(n, struct vmap_area, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3124) if (tmp->va_start <= addr) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3125) va = tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3126) if (tmp->va_end >= addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3127) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3128)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3129) n = n->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3130) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3131) n = n->rb_left;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3132) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3133) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3135) return va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3136) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3137)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3138) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3139) * pvm_determine_end_from_reverse - find the highest aligned address
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3140) * of free block below VMALLOC_END
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3141) * @va:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3142) * in - the VA we start the search(reverse order);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3143) * out - the VA with the highest aligned end address.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3144) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3145) * Returns: determined end address within vmap_area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3146) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3147) static unsigned long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3148) pvm_determine_end_from_reverse(struct vmap_area **va, unsigned long align)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3149) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3150) unsigned long vmalloc_end = VMALLOC_END & ~(align - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3151) unsigned long addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3152)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3153) if (likely(*va)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3154) list_for_each_entry_from_reverse((*va),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3155) &free_vmap_area_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3156) addr = min((*va)->va_end & ~(align - 1), vmalloc_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3157) if ((*va)->va_start < addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3158) return addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3159) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3160) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3161)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3162) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3163) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3164)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3165) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3166) * pcpu_get_vm_areas - allocate vmalloc areas for percpu allocator
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3167) * @offsets: array containing offset of each area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3168) * @sizes: array containing size of each area
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3169) * @nr_vms: the number of areas to allocate
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3170) * @align: alignment, all entries in @offsets and @sizes must be aligned to this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3171) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3172) * Returns: kmalloc'd vm_struct pointer array pointing to allocated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3173) * vm_structs on success, %NULL on failure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3174) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3175) * Percpu allocator wants to use congruent vm areas so that it can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3176) * maintain the offsets among percpu areas. This function allocates
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3177) * congruent vmalloc areas for it with GFP_KERNEL. These areas tend to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3178) * be scattered pretty far, distance between two areas easily going up
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3179) * to gigabytes. To avoid interacting with regular vmallocs, these
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3180) * areas are allocated from top.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3181) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3182) * Despite its complicated look, this allocator is rather simple. It
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3183) * does everything top-down and scans free blocks from the end looking
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3184) * for matching base. While scanning, if any of the areas do not fit the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3185) * base address is pulled down to fit the area. Scanning is repeated till
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3186) * all the areas fit and then all necessary data structures are inserted
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3187) * and the result is returned.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3188) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3189) struct vm_struct **pcpu_get_vm_areas(const unsigned long *offsets,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3190) const size_t *sizes, int nr_vms,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3191) size_t align)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3192) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3193) const unsigned long vmalloc_start = ALIGN(VMALLOC_START, align);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3194) const unsigned long vmalloc_end = VMALLOC_END & ~(align - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3195) struct vmap_area **vas, *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3196) struct vm_struct **vms;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3197) int area, area2, last_area, term_area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3198) unsigned long base, start, size, end, last_end, orig_start, orig_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3199) bool purged = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3200) enum fit_type type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3201)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3202) /* verify parameters and allocate data structures */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3203) BUG_ON(offset_in_page(align) || !is_power_of_2(align));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3204) for (last_area = 0, area = 0; area < nr_vms; area++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3205) start = offsets[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3206) end = start + sizes[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3208) /* is everything aligned properly? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3209) BUG_ON(!IS_ALIGNED(offsets[area], align));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3210) BUG_ON(!IS_ALIGNED(sizes[area], align));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3211)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3212) /* detect the area with the highest address */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3213) if (start > offsets[last_area])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3214) last_area = area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3215)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3216) for (area2 = area + 1; area2 < nr_vms; area2++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3217) unsigned long start2 = offsets[area2];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3218) unsigned long end2 = start2 + sizes[area2];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3220) BUG_ON(start2 < end && start < end2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3221) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3222) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3223) last_end = offsets[last_area] + sizes[last_area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3224)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3225) if (vmalloc_end - vmalloc_start < last_end) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3226) WARN_ON(true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3227) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3228) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3229)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3230) vms = kcalloc(nr_vms, sizeof(vms[0]), GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3231) vas = kcalloc(nr_vms, sizeof(vas[0]), GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3232) if (!vas || !vms)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3233) goto err_free2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3234)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3235) for (area = 0; area < nr_vms; area++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3236) vas[area] = kmem_cache_zalloc(vmap_area_cachep, GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3237) vms[area] = kzalloc(sizeof(struct vm_struct), GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3238) if (!vas[area] || !vms[area])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3239) goto err_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3240) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3241) retry:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3242) spin_lock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3243)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3244) /* start scanning - we scan from the top, begin with the last area */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3245) area = term_area = last_area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3246) start = offsets[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3247) end = start + sizes[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3248)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3249) va = pvm_find_va_enclose_addr(vmalloc_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3250) base = pvm_determine_end_from_reverse(&va, align) - end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3251)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3252) while (true) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3253) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3254) * base might have underflowed, add last_end before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3255) * comparing.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3256) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3257) if (base + last_end < vmalloc_start + last_end)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3258) goto overflow;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3259)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3260) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3261) * Fitting base has not been found.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3262) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3263) if (va == NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3264) goto overflow;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3265)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3266) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3267) * If required width exceeds current VA block, move
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3268) * base downwards and then recheck.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3269) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3270) if (base + end > va->va_end) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3271) base = pvm_determine_end_from_reverse(&va, align) - end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3272) term_area = area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3273) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3274) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3275)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3276) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3277) * If this VA does not fit, move base downwards and recheck.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3278) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3279) if (base + start < va->va_start) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3280) va = node_to_va(rb_prev(&va->rb_node));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3281) base = pvm_determine_end_from_reverse(&va, align) - end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3282) term_area = area;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3283) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3284) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3285)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3286) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3287) * This area fits, move on to the previous one. If
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3288) * the previous one is the terminal one, we're done.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3289) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3290) area = (area + nr_vms - 1) % nr_vms;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3291) if (area == term_area)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3292) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3293)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3294) start = offsets[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3295) end = start + sizes[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3296) va = pvm_find_va_enclose_addr(base + end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3297) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3298)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3299) /* we've found a fitting base, insert all va's */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3300) for (area = 0; area < nr_vms; area++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3301) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3302)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3303) start = base + offsets[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3304) size = sizes[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3305)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3306) va = pvm_find_va_enclose_addr(start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3307) if (WARN_ON_ONCE(va == NULL))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3308) /* It is a BUG(), but trigger recovery instead. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3309) goto recovery;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3310)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3311) type = classify_va_fit_type(va, start, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3312) if (WARN_ON_ONCE(type == NOTHING_FIT))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3313) /* It is a BUG(), but trigger recovery instead. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3314) goto recovery;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3315)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3316) ret = adjust_va_to_fit_type(va, start, size, type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3317) if (unlikely(ret))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3318) goto recovery;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3319)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3320) /* Allocated area. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3321) va = vas[area];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3322) va->va_start = start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3323) va->va_end = start + size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3324) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3325)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3326) spin_unlock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3327)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3328) /* populate the kasan shadow space */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3329) for (area = 0; area < nr_vms; area++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3330) if (kasan_populate_vmalloc(vas[area]->va_start, sizes[area]))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3331) goto err_free_shadow;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3332)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3333) kasan_unpoison_vmalloc((void *)vas[area]->va_start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3334) sizes[area]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3335) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3336)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3337) /* insert all vm's */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3338) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3339) for (area = 0; area < nr_vms; area++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3340) insert_vmap_area(vas[area], &vmap_area_root, &vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3341)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3342) setup_vmalloc_vm_locked(vms[area], vas[area], VM_ALLOC,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3343) pcpu_get_vm_areas);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3344) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3345) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3346)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3347) kfree(vas);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3348) return vms;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3349)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3350) recovery:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3351) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3352) * Remove previously allocated areas. There is no
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3353) * need in removing these areas from the busy tree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3354) * because they are inserted only on the final step
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3355) * and when pcpu_get_vm_areas() is success.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3356) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3357) while (area--) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3358) orig_start = vas[area]->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3359) orig_end = vas[area]->va_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3360) va = merge_or_add_vmap_area(vas[area], &free_vmap_area_root,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3361) &free_vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3362) if (va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3363) kasan_release_vmalloc(orig_start, orig_end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3364) va->va_start, va->va_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3365) vas[area] = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3366) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3367)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3368) overflow:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3369) spin_unlock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3370) if (!purged) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3371) purge_vmap_area_lazy();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3372) purged = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3373)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3374) /* Before "retry", check if we recover. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3375) for (area = 0; area < nr_vms; area++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3376) if (vas[area])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3377) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3378)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3379) vas[area] = kmem_cache_zalloc(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3380) vmap_area_cachep, GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3381) if (!vas[area])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3382) goto err_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3383) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3384)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3385) goto retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3386) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3387)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3388) err_free:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3389) for (area = 0; area < nr_vms; area++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3390) if (vas[area])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3391) kmem_cache_free(vmap_area_cachep, vas[area]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3392)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3393) kfree(vms[area]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3394) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3395) err_free2:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3396) kfree(vas);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3397) kfree(vms);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3398) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3399)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3400) err_free_shadow:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3401) spin_lock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3402) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3403) * We release all the vmalloc shadows, even the ones for regions that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3404) * hadn't been successfully added. This relies on kasan_release_vmalloc
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3405) * being able to tolerate this case.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3406) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3407) for (area = 0; area < nr_vms; area++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3408) orig_start = vas[area]->va_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3409) orig_end = vas[area]->va_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3410) va = merge_or_add_vmap_area(vas[area], &free_vmap_area_root,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3411) &free_vmap_area_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3412) if (va)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3413) kasan_release_vmalloc(orig_start, orig_end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3414) va->va_start, va->va_end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3415) vas[area] = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3416) kfree(vms[area]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3417) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3418) spin_unlock(&free_vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3419) kfree(vas);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3420) kfree(vms);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3421) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3422) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3423)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3424) /**
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3425) * pcpu_free_vm_areas - free vmalloc areas for percpu allocator
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3426) * @vms: vm_struct pointer array returned by pcpu_get_vm_areas()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3427) * @nr_vms: the number of allocated areas
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3428) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3429) * Free vm_structs and the array allocated by pcpu_get_vm_areas().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3430) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3431) void pcpu_free_vm_areas(struct vm_struct **vms, int nr_vms)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3432) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3433) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3434)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3435) for (i = 0; i < nr_vms; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3436) free_vm_area(vms[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3437) kfree(vms);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3438) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3439) #endif /* CONFIG_SMP */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3440)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3441) #ifdef CONFIG_PROC_FS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3442) static void *s_start(struct seq_file *m, loff_t *pos)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3443) __acquires(&vmap_purge_lock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3444) __acquires(&vmap_area_lock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3445) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3446) mutex_lock(&vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3447) spin_lock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3448)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3449) return seq_list_start(&vmap_area_list, *pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3450) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3451)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3452) static void *s_next(struct seq_file *m, void *p, loff_t *pos)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3453) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3454) return seq_list_next(p, &vmap_area_list, pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3455) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3456)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3457) static void s_stop(struct seq_file *m, void *p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3458) __releases(&vmap_area_lock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3459) __releases(&vmap_purge_lock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3460) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3461) spin_unlock(&vmap_area_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3462) mutex_unlock(&vmap_purge_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3463) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3464)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3465) static void show_numa_info(struct seq_file *m, struct vm_struct *v)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3466) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3467) if (IS_ENABLED(CONFIG_NUMA)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3468) unsigned int nr, *counters = m->private;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3469)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3470) if (!counters)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3471) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3472)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3473) if (v->flags & VM_UNINITIALIZED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3474) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3475) /* Pair with smp_wmb() in clear_vm_uninitialized_flag() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3476) smp_rmb();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3477)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3478) memset(counters, 0, nr_node_ids * sizeof(unsigned int));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3479)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3480) for (nr = 0; nr < v->nr_pages; nr++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3481) counters[page_to_nid(v->pages[nr])]++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3482)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3483) for_each_node_state(nr, N_HIGH_MEMORY)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3484) if (counters[nr])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3485) seq_printf(m, " N%u=%u", nr, counters[nr]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3486) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3487) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3488)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3489) static void show_purge_info(struct seq_file *m)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3490) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3491) struct llist_node *head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3492) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3493)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3494) head = READ_ONCE(vmap_purge_list.first);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3495) if (head == NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3496) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3497)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3498) llist_for_each_entry(va, head, purge_list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3499) seq_printf(m, "0x%pK-0x%pK %7ld unpurged vm_area\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3500) (void *)va->va_start, (void *)va->va_end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3501) va->va_end - va->va_start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3502) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3503) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3504)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3505) static int s_show(struct seq_file *m, void *p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3506) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3507) struct vmap_area *va;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3508) struct vm_struct *v;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3509)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3510) va = list_entry(p, struct vmap_area, list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3511)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3512) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3513) * s_show can encounter race with remove_vm_area, !vm on behalf
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3514) * of vmap area is being tear down or vm_map_ram allocation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3515) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3516) if (!va->vm) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3517) seq_printf(m, "0x%pK-0x%pK %7ld vm_map_ram\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3518) (void *)va->va_start, (void *)va->va_end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3519) va->va_end - va->va_start);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3520)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3521) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3522) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3523)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3524) v = va->vm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3525)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3526) seq_printf(m, "0x%pK-0x%pK %7ld",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3527) v->addr, v->addr + v->size, v->size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3528)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3529) if (v->caller)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3530) seq_printf(m, " %pS", v->caller);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3531)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3532) if (v->nr_pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3533) seq_printf(m, " pages=%d", v->nr_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3534)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3535) if (v->phys_addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3536) seq_printf(m, " phys=%pa", &v->phys_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3537)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3538) if (v->flags & VM_IOREMAP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3539) seq_puts(m, " ioremap");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3540)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3541) if (v->flags & VM_ALLOC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3542) seq_puts(m, " vmalloc");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3543)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3544) if (v->flags & VM_MAP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3545) seq_puts(m, " vmap");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3546)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3547) if (v->flags & VM_USERMAP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3548) seq_puts(m, " user");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3549)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3550) if (v->flags & VM_DMA_COHERENT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3551) seq_puts(m, " dma-coherent");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3552)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3553) if (is_vmalloc_addr(v->pages))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3554) seq_puts(m, " vpages");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3555)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3556) show_numa_info(m, v);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3557) trace_android_vh_show_stack_hash(m, v);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3558) seq_putc(m, '\n');
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3559)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3560) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3561) * As a final step, dump "unpurged" areas. Note,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3562) * that entire "/proc/vmallocinfo" output will not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3563) * be address sorted, because the purge list is not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3564) * sorted.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3565) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3566) if (list_is_last(&va->list, &vmap_area_list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3567) show_purge_info(m);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3568)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3569) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3570) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3571)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3572) static const struct seq_operations vmalloc_op = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3573) .start = s_start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3574) .next = s_next,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3575) .stop = s_stop,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3576) .show = s_show,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3577) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3578)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3579) static int __init proc_vmalloc_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3580) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3581) if (IS_ENABLED(CONFIG_NUMA))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3582) proc_create_seq_private("vmallocinfo", 0400, NULL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3583) &vmalloc_op,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3584) nr_node_ids * sizeof(unsigned int), NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3585) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3586) proc_create_seq("vmallocinfo", 0400, NULL, &vmalloc_op);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3587) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3588) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3589) module_init(proc_vmalloc_init);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3590)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3591) #endif
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags