^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * IBM System z Huge TLB Page Support for Kernel.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * Copyright IBM Corp. 2007,2020
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) * Author(s): Gerald Schaefer <gerald.schaefer@de.ibm.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #define KMSG_COMPONENT "hugetlb"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11)
^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/hugetlb.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <linux/mman.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <linux/sched/mm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include <linux/security.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) * If the bit selected by single-bit bitmask "a" is set within "x", move
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) * it to the position indicated by single-bit bitmask "b".
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #define move_set_bit(x, a, b) (((x) & (a)) >> ilog2(a) << ilog2(b))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) static inline unsigned long __pte_to_rste(pte_t pte)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) unsigned long rste;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) * Convert encoding pte bits pmd / pud bits
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) * lIR.uswrdy.p dy..R...I...wr
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) * empty 010.000000.0 -> 00..0...1...00
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) * prot-none, clean, old 111.000000.1 -> 00..1...1...00
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) * prot-none, clean, young 111.000001.1 -> 01..1...1...00
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) * prot-none, dirty, old 111.000010.1 -> 10..1...1...00
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) * prot-none, dirty, young 111.000011.1 -> 11..1...1...00
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) * read-only, clean, old 111.000100.1 -> 00..1...1...01
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) * read-only, clean, young 101.000101.1 -> 01..1...0...01
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) * read-only, dirty, old 111.000110.1 -> 10..1...1...01
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) * read-only, dirty, young 101.000111.1 -> 11..1...0...01
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) * read-write, clean, old 111.001100.1 -> 00..1...1...11
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) * read-write, clean, young 101.001101.1 -> 01..1...0...11
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) * read-write, dirty, old 110.001110.1 -> 10..0...1...11
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) * read-write, dirty, young 100.001111.1 -> 11..0...0...11
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) * HW-bits: R read-only, I invalid
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) * SW-bits: p present, y young, d dirty, r read, w write, s special,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) * u unused, l large
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) if (pte_present(pte)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) rste = pte_val(pte) & PAGE_MASK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) rste |= move_set_bit(pte_val(pte), _PAGE_READ,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) _SEGMENT_ENTRY_READ);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) rste |= move_set_bit(pte_val(pte), _PAGE_WRITE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) _SEGMENT_ENTRY_WRITE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) rste |= move_set_bit(pte_val(pte), _PAGE_INVALID,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) _SEGMENT_ENTRY_INVALID);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) rste |= move_set_bit(pte_val(pte), _PAGE_PROTECT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) _SEGMENT_ENTRY_PROTECT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) rste |= move_set_bit(pte_val(pte), _PAGE_DIRTY,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) _SEGMENT_ENTRY_DIRTY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) rste |= move_set_bit(pte_val(pte), _PAGE_YOUNG,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) _SEGMENT_ENTRY_YOUNG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) #ifdef CONFIG_MEM_SOFT_DIRTY
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) rste |= move_set_bit(pte_val(pte), _PAGE_SOFT_DIRTY,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) _SEGMENT_ENTRY_SOFT_DIRTY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) rste |= move_set_bit(pte_val(pte), _PAGE_NOEXEC,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) _SEGMENT_ENTRY_NOEXEC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) rste = _SEGMENT_ENTRY_EMPTY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) return rste;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) static inline pte_t __rste_to_pte(unsigned long rste)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) int present;
^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) if ((rste & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) present = pud_present(__pud(rste));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) present = pmd_present(__pmd(rste));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) * Convert encoding pmd / pud bits pte bits
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) * dy..R...I...wr lIR.uswrdy.p
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) * empty 00..0...1...00 -> 010.000000.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) * prot-none, clean, old 00..1...1...00 -> 111.000000.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) * prot-none, clean, young 01..1...1...00 -> 111.000001.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) * prot-none, dirty, old 10..1...1...00 -> 111.000010.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) * prot-none, dirty, young 11..1...1...00 -> 111.000011.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) * read-only, clean, old 00..1...1...01 -> 111.000100.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) * read-only, clean, young 01..1...0...01 -> 101.000101.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) * read-only, dirty, old 10..1...1...01 -> 111.000110.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) * read-only, dirty, young 11..1...0...01 -> 101.000111.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) * read-write, clean, old 00..1...1...11 -> 111.001100.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) * read-write, clean, young 01..1...0...11 -> 101.001101.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) * read-write, dirty, old 10..0...1...11 -> 110.001110.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) * read-write, dirty, young 11..0...0...11 -> 100.001111.1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) * HW-bits: R read-only, I invalid
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) * SW-bits: p present, y young, d dirty, r read, w write, s special,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) * u unused, l large
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) if (present) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) pte_val(pte) = rste & _SEGMENT_ENTRY_ORIGIN_LARGE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) pte_val(pte) |= _PAGE_LARGE | _PAGE_PRESENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) pte_val(pte) |= move_set_bit(rste, _SEGMENT_ENTRY_READ,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) _PAGE_READ);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) pte_val(pte) |= move_set_bit(rste, _SEGMENT_ENTRY_WRITE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) _PAGE_WRITE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) pte_val(pte) |= move_set_bit(rste, _SEGMENT_ENTRY_INVALID,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) _PAGE_INVALID);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) pte_val(pte) |= move_set_bit(rste, _SEGMENT_ENTRY_PROTECT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) _PAGE_PROTECT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) pte_val(pte) |= move_set_bit(rste, _SEGMENT_ENTRY_DIRTY,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) _PAGE_DIRTY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) pte_val(pte) |= move_set_bit(rste, _SEGMENT_ENTRY_YOUNG,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) _PAGE_YOUNG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) #ifdef CONFIG_MEM_SOFT_DIRTY
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) pte_val(pte) |= move_set_bit(rste, _SEGMENT_ENTRY_SOFT_DIRTY,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) _PAGE_SOFT_DIRTY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) pte_val(pte) |= move_set_bit(rste, _SEGMENT_ENTRY_NOEXEC,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) _PAGE_NOEXEC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) pte_val(pte) = _PAGE_INVALID;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) return pte;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) static void clear_huge_pte_skeys(struct mm_struct *mm, unsigned long rste)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) unsigned long size, paddr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) if (!mm_uses_skeys(mm) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) rste & _SEGMENT_ENTRY_INVALID)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) if ((rste & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) page = pud_page(__pud(rste));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) size = PUD_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) paddr = rste & PUD_MASK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) page = pmd_page(__pmd(rste));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) size = PMD_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) paddr = rste & PMD_MASK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) if (!test_and_set_bit(PG_arch_1, &page->flags))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) __storage_key_init_range(paddr, paddr + size - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) pte_t *ptep, pte_t pte)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) unsigned long rste;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) rste = __pte_to_rste(pte);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) if (!MACHINE_HAS_NX)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) rste &= ~_SEGMENT_ENTRY_NOEXEC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) /* Set correct table type for 2G hugepages */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) if ((pte_val(*ptep) & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) if (likely(pte_present(pte)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) rste |= _REGION3_ENTRY_LARGE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) rste |= _REGION_ENTRY_TYPE_R3;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) } else if (likely(pte_present(pte)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) rste |= _SEGMENT_ENTRY_LARGE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) clear_huge_pte_skeys(mm, rste);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) pte_val(*ptep) = rste;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) pte_t huge_ptep_get(pte_t *ptep)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) return __rste_to_pte(pte_val(*ptep));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) pte_t huge_ptep_get_and_clear(struct mm_struct *mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) unsigned long addr, pte_t *ptep)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) pte_t pte = huge_ptep_get(ptep);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) pmd_t *pmdp = (pmd_t *) ptep;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) pud_t *pudp = (pud_t *) ptep;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) if ((pte_val(*ptep) & _REGION_ENTRY_TYPE_MASK) == _REGION_ENTRY_TYPE_R3)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) pudp_xchg_direct(mm, addr, pudp, __pud(_REGION3_ENTRY_EMPTY));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) pmdp_xchg_direct(mm, addr, pmdp, __pmd(_SEGMENT_ENTRY_EMPTY));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) return pte;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) pte_t *huge_pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) unsigned long addr, unsigned long sz)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) pgd_t *pgdp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) p4d_t *p4dp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) pud_t *pudp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) pmd_t *pmdp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) pgdp = pgd_offset(mm, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) p4dp = p4d_alloc(mm, pgdp, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) if (p4dp) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) pudp = pud_alloc(mm, p4dp, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) if (pudp) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) if (sz == PUD_SIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) return (pte_t *) pudp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) else if (sz == PMD_SIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) pmdp = pmd_alloc(mm, pudp, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) return (pte_t *) pmdp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) pte_t *huge_pte_offset(struct mm_struct *mm,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) unsigned long addr, unsigned long sz)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) pgd_t *pgdp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) p4d_t *p4dp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) pud_t *pudp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) pmd_t *pmdp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) pgdp = pgd_offset(mm, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) if (pgd_present(*pgdp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) p4dp = p4d_offset(pgdp, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) if (p4d_present(*p4dp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) pudp = pud_offset(p4dp, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) if (pud_present(*pudp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) if (pud_large(*pudp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) return (pte_t *) pudp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) pmdp = pmd_offset(pudp, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) return (pte_t *) pmdp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) int pmd_huge(pmd_t pmd)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) return pmd_large(pmd);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) int pud_huge(pud_t pud)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) return pud_large(pud);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) struct page *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) follow_huge_pud(struct mm_struct *mm, unsigned long address,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) pud_t *pud, int flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) if (flags & FOLL_GET)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) return pud_page(*pud) + ((address & ~PUD_MASK) >> PAGE_SHIFT);
^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) bool __init arch_hugetlb_valid_size(unsigned long size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) if (MACHINE_HAS_EDAT1 && size == PMD_SIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) else if (MACHINE_HAS_EDAT2 && size == PUD_SIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) static unsigned long hugetlb_get_unmapped_area_bottomup(struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) unsigned long addr, unsigned long len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) unsigned long pgoff, unsigned long flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) struct hstate *h = hstate_file(file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) struct vm_unmapped_area_info info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) info.flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) info.length = len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) info.low_limit = current->mm->mmap_base;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) info.high_limit = TASK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) info.align_mask = PAGE_MASK & ~huge_page_mask(h);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) info.align_offset = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) return vm_unmapped_area(&info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) static unsigned long hugetlb_get_unmapped_area_topdown(struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) unsigned long addr0, unsigned long len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) unsigned long pgoff, unsigned long flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) struct hstate *h = hstate_file(file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) struct vm_unmapped_area_info info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) unsigned long addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) info.flags = VM_UNMAPPED_AREA_TOPDOWN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) info.length = len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) info.low_limit = max(PAGE_SIZE, mmap_min_addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) info.high_limit = current->mm->mmap_base;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) info.align_mask = PAGE_MASK & ~huge_page_mask(h);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) info.align_offset = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) addr = vm_unmapped_area(&info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) * A failed mmap() very likely causes application failure,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) * so fall back to the bottom-up function here. This scenario
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) * can happen with large stack limits and large mmap()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) * allocations.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) if (addr & ~PAGE_MASK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) VM_BUG_ON(addr != -ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) info.flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) info.low_limit = TASK_UNMAPPED_BASE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) info.high_limit = TASK_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) addr = vm_unmapped_area(&info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) return addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) unsigned long len, unsigned long pgoff, unsigned long flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) struct hstate *h = hstate_file(file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) struct mm_struct *mm = current->mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) struct vm_area_struct *vma;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) if (len & ~huge_page_mask(h))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) if (len > TASK_SIZE - mmap_min_addr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) if (flags & MAP_FIXED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) if (prepare_hugepage_range(file, addr, len))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) goto check_asce_limit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) if (addr) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) addr = ALIGN(addr, huge_page_size(h));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) vma = find_vma(mm, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) if (TASK_SIZE - len >= addr && addr >= mmap_min_addr &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) (!vma || addr + len <= vm_start_gap(vma)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) goto check_asce_limit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) if (mm->get_unmapped_area == arch_get_unmapped_area)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) addr = hugetlb_get_unmapped_area_bottomup(file, addr, len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) pgoff, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) addr = hugetlb_get_unmapped_area_topdown(file, addr, len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) pgoff, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) if (offset_in_page(addr))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) return addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) check_asce_limit:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) return check_asce_limit(mm, addr, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) }