Loading...
1// SPDX-License-Identifier: GPL-2.0
2/*
3 * This file implements KASLR memory randomization for x86_64. It randomizes
4 * the virtual address space of kernel memory regions (physical memory
5 * mapping, vmalloc & vmemmap) for x86_64. This security feature mitigates
6 * exploits relying on predictable kernel addresses.
7 *
8 * Entropy is generated using the KASLR early boot functions now shared in
9 * the lib directory (originally written by Kees Cook). Randomization is
10 * done on PGD & P4D/PUD page table levels to increase possible addresses.
11 * The physical memory mapping code was adapted to support P4D/PUD level
12 * virtual addresses. This implementation on the best configuration provides
13 * 30,000 possible virtual addresses in average for each memory region.
14 * An additional low memory page is used to ensure each CPU can start with
15 * a PGD aligned virtual address (for realmode).
16 *
17 * The order of each memory region is not changed. The feature looks at
18 * the available space for the regions based on different configuration
19 * options and randomizes the base and space between each. The size of the
20 * physical memory mapping is the available physical memory.
21 */
22
23#include <linux/kernel.h>
24#include <linux/init.h>
25#include <linux/random.h>
26#include <linux/memblock.h>
27
28#include <asm/pgalloc.h>
29#include <asm/pgtable.h>
30#include <asm/setup.h>
31#include <asm/kaslr.h>
32
33#include "mm_internal.h"
34
35#define TB_SHIFT 40
36
37/*
38 * The end address could depend on more configuration options to make the
39 * highest amount of space for randomization available, but that's too hard
40 * to keep straight and caused issues already.
41 */
42static const unsigned long vaddr_end = CPU_ENTRY_AREA_BASE;
43
44/*
45 * Memory regions randomized by KASLR (except modules that use a separate logic
46 * earlier during boot). The list is ordered based on virtual addresses. This
47 * order is kept after randomization.
48 */
49static __initdata struct kaslr_memory_region {
50 unsigned long *base;
51 unsigned long size_tb;
52} kaslr_regions[] = {
53 { &page_offset_base, 0 },
54 { &vmalloc_base, 0 },
55 { &vmemmap_base, 0 },
56};
57
58/* Get size in bytes used by the memory region */
59static inline unsigned long get_padding(struct kaslr_memory_region *region)
60{
61 return (region->size_tb << TB_SHIFT);
62}
63
64/*
65 * Apply no randomization if KASLR was disabled at boot or if KASAN
66 * is enabled. KASAN shadow mappings rely on regions being PGD aligned.
67 */
68static inline bool kaslr_memory_enabled(void)
69{
70 return kaslr_enabled() && !IS_ENABLED(CONFIG_KASAN);
71}
72
73/* Initialize base and padding for each memory region randomized with KASLR */
74void __init kernel_randomize_memory(void)
75{
76 size_t i;
77 unsigned long vaddr_start, vaddr;
78 unsigned long rand, memory_tb;
79 struct rnd_state rand_state;
80 unsigned long remain_entropy;
81 unsigned long vmemmap_size;
82
83 vaddr_start = pgtable_l5_enabled() ? __PAGE_OFFSET_BASE_L5 : __PAGE_OFFSET_BASE_L4;
84 vaddr = vaddr_start;
85
86 /*
87 * These BUILD_BUG_ON checks ensure the memory layout is consistent
88 * with the vaddr_start/vaddr_end variables. These checks are very
89 * limited....
90 */
91 BUILD_BUG_ON(vaddr_start >= vaddr_end);
92 BUILD_BUG_ON(vaddr_end != CPU_ENTRY_AREA_BASE);
93 BUILD_BUG_ON(vaddr_end > __START_KERNEL_map);
94
95 if (!kaslr_memory_enabled())
96 return;
97
98 kaslr_regions[0].size_tb = 1 << (MAX_PHYSMEM_BITS - TB_SHIFT);
99 kaslr_regions[1].size_tb = VMALLOC_SIZE_TB;
100
101 /*
102 * Update Physical memory mapping to available and
103 * add padding if needed (especially for memory hotplug support).
104 */
105 BUG_ON(kaslr_regions[0].base != &page_offset_base);
106 memory_tb = DIV_ROUND_UP(max_pfn << PAGE_SHIFT, 1UL << TB_SHIFT) +
107 CONFIG_RANDOMIZE_MEMORY_PHYSICAL_PADDING;
108
109 /* Adapt phyiscal memory region size based on available memory */
110 if (memory_tb < kaslr_regions[0].size_tb)
111 kaslr_regions[0].size_tb = memory_tb;
112
113 /*
114 * Calculate the vmemmap region size in TBs, aligned to a TB
115 * boundary.
116 */
117 vmemmap_size = (kaslr_regions[0].size_tb << (TB_SHIFT - PAGE_SHIFT)) *
118 sizeof(struct page);
119 kaslr_regions[2].size_tb = DIV_ROUND_UP(vmemmap_size, 1UL << TB_SHIFT);
120
121 /* Calculate entropy available between regions */
122 remain_entropy = vaddr_end - vaddr_start;
123 for (i = 0; i < ARRAY_SIZE(kaslr_regions); i++)
124 remain_entropy -= get_padding(&kaslr_regions[i]);
125
126 prandom_seed_state(&rand_state, kaslr_get_random_long("Memory"));
127
128 for (i = 0; i < ARRAY_SIZE(kaslr_regions); i++) {
129 unsigned long entropy;
130
131 /*
132 * Select a random virtual address using the extra entropy
133 * available.
134 */
135 entropy = remain_entropy / (ARRAY_SIZE(kaslr_regions) - i);
136 prandom_bytes_state(&rand_state, &rand, sizeof(rand));
137 entropy = (rand % (entropy + 1)) & PUD_MASK;
138 vaddr += entropy;
139 *kaslr_regions[i].base = vaddr;
140
141 /*
142 * Jump the region and add a minimum padding based on
143 * randomization alignment.
144 */
145 vaddr += get_padding(&kaslr_regions[i]);
146 vaddr = round_up(vaddr + 1, PUD_SIZE);
147 remain_entropy -= entropy;
148 }
149}
150
151static void __meminit init_trampoline_pud(void)
152{
153 pud_t *pud_page_tramp, *pud, *pud_tramp;
154 p4d_t *p4d_page_tramp, *p4d, *p4d_tramp;
155 unsigned long paddr, vaddr;
156 pgd_t *pgd;
157
158 pud_page_tramp = alloc_low_page();
159
160 /*
161 * There are two mappings for the low 1MB area, the direct mapping
162 * and the 1:1 mapping for the real mode trampoline:
163 *
164 * Direct mapping: virt_addr = phys_addr + PAGE_OFFSET
165 * 1:1 mapping: virt_addr = phys_addr
166 */
167 paddr = 0;
168 vaddr = (unsigned long)__va(paddr);
169 pgd = pgd_offset_k(vaddr);
170
171 p4d = p4d_offset(pgd, vaddr);
172 pud = pud_offset(p4d, vaddr);
173
174 pud_tramp = pud_page_tramp + pud_index(paddr);
175 *pud_tramp = *pud;
176
177 if (pgtable_l5_enabled()) {
178 p4d_page_tramp = alloc_low_page();
179
180 p4d_tramp = p4d_page_tramp + p4d_index(paddr);
181
182 set_p4d(p4d_tramp,
183 __p4d(_KERNPG_TABLE | __pa(pud_page_tramp)));
184
185 set_pgd(&trampoline_pgd_entry,
186 __pgd(_KERNPG_TABLE | __pa(p4d_page_tramp)));
187 } else {
188 set_pgd(&trampoline_pgd_entry,
189 __pgd(_KERNPG_TABLE | __pa(pud_page_tramp)));
190 }
191}
192
193/*
194 * The real mode trampoline, which is required for bootstrapping CPUs
195 * occupies only a small area under the low 1MB. See reserve_real_mode()
196 * for details.
197 *
198 * If KASLR is disabled the first PGD entry of the direct mapping is copied
199 * to map the real mode trampoline.
200 *
201 * If KASLR is enabled, copy only the PUD which covers the low 1MB
202 * area. This limits the randomization granularity to 1GB for both 4-level
203 * and 5-level paging.
204 */
205void __meminit init_trampoline(void)
206{
207 if (!kaslr_memory_enabled()) {
208 init_trampoline_default();
209 return;
210 }
211
212 init_trampoline_pud();
213}
1// SPDX-License-Identifier: GPL-2.0
2/*
3 * This file implements KASLR memory randomization for x86_64. It randomizes
4 * the virtual address space of kernel memory regions (physical memory
5 * mapping, vmalloc & vmemmap) for x86_64. This security feature mitigates
6 * exploits relying on predictable kernel addresses.
7 *
8 * Entropy is generated using the KASLR early boot functions now shared in
9 * the lib directory (originally written by Kees Cook). Randomization is
10 * done on PGD & P4D/PUD page table levels to increase possible addresses.
11 * The physical memory mapping code was adapted to support P4D/PUD level
12 * virtual addresses. This implementation on the best configuration provides
13 * 30,000 possible virtual addresses in average for each memory region.
14 * An additional low memory page is used to ensure each CPU can start with
15 * a PGD aligned virtual address (for realmode).
16 *
17 * The order of each memory region is not changed. The feature looks at
18 * the available space for the regions based on different configuration
19 * options and randomizes the base and space between each. The size of the
20 * physical memory mapping is the available physical memory.
21 */
22
23#include <linux/kernel.h>
24#include <linux/init.h>
25#include <linux/prandom.h>
26#include <linux/memblock.h>
27#include <linux/pgtable.h>
28
29#include <asm/setup.h>
30#include <asm/kaslr.h>
31
32#include "mm_internal.h"
33
34#define TB_SHIFT 40
35
36/*
37 * The end address could depend on more configuration options to make the
38 * highest amount of space for randomization available, but that's too hard
39 * to keep straight and caused issues already.
40 */
41static const unsigned long vaddr_end = CPU_ENTRY_AREA_BASE;
42
43/*
44 * Memory regions randomized by KASLR (except modules that use a separate logic
45 * earlier during boot). The list is ordered based on virtual addresses. This
46 * order is kept after randomization.
47 */
48static __initdata struct kaslr_memory_region {
49 unsigned long *base;
50 unsigned long *end;
51 unsigned long size_tb;
52} kaslr_regions[] = {
53 {
54 .base = &page_offset_base,
55 .end = &direct_map_physmem_end,
56 },
57 {
58 .base = &vmalloc_base,
59 },
60 {
61 .base = &vmemmap_base,
62 },
63};
64
65/*
66 * The end of the physical address space that can be mapped directly by the
67 * kernel. This starts out at (1<<MAX_PHYSMEM_BITS) - 1), but KASLR may reduce
68 * that in order to increase the available entropy for mapping other regions.
69 */
70unsigned long direct_map_physmem_end __ro_after_init;
71
72/* Get size in bytes used by the memory region */
73static inline unsigned long get_padding(struct kaslr_memory_region *region)
74{
75 return (region->size_tb << TB_SHIFT);
76}
77
78/* Initialize base and padding for each memory region randomized with KASLR */
79void __init kernel_randomize_memory(void)
80{
81 size_t i;
82 unsigned long vaddr_start, vaddr;
83 unsigned long rand, memory_tb;
84 struct rnd_state rand_state;
85 unsigned long remain_entropy;
86 unsigned long vmemmap_size;
87
88 vaddr_start = pgtable_l5_enabled() ? __PAGE_OFFSET_BASE_L5 : __PAGE_OFFSET_BASE_L4;
89 vaddr = vaddr_start;
90
91 /*
92 * These BUILD_BUG_ON checks ensure the memory layout is consistent
93 * with the vaddr_start/vaddr_end variables. These checks are very
94 * limited....
95 */
96 BUILD_BUG_ON(vaddr_start >= vaddr_end);
97 BUILD_BUG_ON(vaddr_end != CPU_ENTRY_AREA_BASE);
98 BUILD_BUG_ON(vaddr_end > __START_KERNEL_map);
99
100 /* Preset the end of the possible address space for physical memory */
101 direct_map_physmem_end = ((1ULL << MAX_PHYSMEM_BITS) - 1);
102 if (!kaslr_memory_enabled())
103 return;
104
105 kaslr_regions[0].size_tb = 1 << (MAX_PHYSMEM_BITS - TB_SHIFT);
106 kaslr_regions[1].size_tb = VMALLOC_SIZE_TB;
107
108 /*
109 * Update Physical memory mapping to available and
110 * add padding if needed (especially for memory hotplug support).
111 */
112 BUG_ON(kaslr_regions[0].base != &page_offset_base);
113 memory_tb = DIV_ROUND_UP(max_pfn << PAGE_SHIFT, 1UL << TB_SHIFT) +
114 CONFIG_RANDOMIZE_MEMORY_PHYSICAL_PADDING;
115
116 /* Adapt physical memory region size based on available memory */
117 if (memory_tb < kaslr_regions[0].size_tb)
118 kaslr_regions[0].size_tb = memory_tb;
119
120 /*
121 * Calculate the vmemmap region size in TBs, aligned to a TB
122 * boundary.
123 */
124 vmemmap_size = (kaslr_regions[0].size_tb << (TB_SHIFT - PAGE_SHIFT)) *
125 sizeof(struct page);
126 kaslr_regions[2].size_tb = DIV_ROUND_UP(vmemmap_size, 1UL << TB_SHIFT);
127
128 /* Calculate entropy available between regions */
129 remain_entropy = vaddr_end - vaddr_start;
130 for (i = 0; i < ARRAY_SIZE(kaslr_regions); i++)
131 remain_entropy -= get_padding(&kaslr_regions[i]);
132
133 prandom_seed_state(&rand_state, kaslr_get_random_long("Memory"));
134
135 for (i = 0; i < ARRAY_SIZE(kaslr_regions); i++) {
136 unsigned long entropy;
137
138 /*
139 * Select a random virtual address using the extra entropy
140 * available.
141 */
142 entropy = remain_entropy / (ARRAY_SIZE(kaslr_regions) - i);
143 prandom_bytes_state(&rand_state, &rand, sizeof(rand));
144 entropy = (rand % (entropy + 1)) & PUD_MASK;
145 vaddr += entropy;
146 *kaslr_regions[i].base = vaddr;
147
148 /* Calculate the end of the region */
149 vaddr += get_padding(&kaslr_regions[i]);
150 /*
151 * KASLR trims the maximum possible size of the
152 * direct-map. Update the direct_map_physmem_end boundary.
153 * No rounding required as the region starts
154 * PUD aligned and size is in units of TB.
155 */
156 if (kaslr_regions[i].end)
157 *kaslr_regions[i].end = __pa_nodebug(vaddr - 1);
158
159 /* Add a minimum padding based on randomization alignment. */
160 vaddr = round_up(vaddr + 1, PUD_SIZE);
161 remain_entropy -= entropy;
162 }
163}
164
165void __meminit init_trampoline_kaslr(void)
166{
167 pud_t *pud_page_tramp, *pud, *pud_tramp;
168 p4d_t *p4d_page_tramp, *p4d, *p4d_tramp;
169 unsigned long paddr, vaddr;
170 pgd_t *pgd;
171
172 pud_page_tramp = alloc_low_page();
173
174 /*
175 * There are two mappings for the low 1MB area, the direct mapping
176 * and the 1:1 mapping for the real mode trampoline:
177 *
178 * Direct mapping: virt_addr = phys_addr + PAGE_OFFSET
179 * 1:1 mapping: virt_addr = phys_addr
180 */
181 paddr = 0;
182 vaddr = (unsigned long)__va(paddr);
183 pgd = pgd_offset_k(vaddr);
184
185 p4d = p4d_offset(pgd, vaddr);
186 pud = pud_offset(p4d, vaddr);
187
188 pud_tramp = pud_page_tramp + pud_index(paddr);
189 *pud_tramp = *pud;
190
191 if (pgtable_l5_enabled()) {
192 p4d_page_tramp = alloc_low_page();
193
194 p4d_tramp = p4d_page_tramp + p4d_index(paddr);
195
196 set_p4d(p4d_tramp,
197 __p4d(_KERNPG_TABLE | __pa(pud_page_tramp)));
198
199 trampoline_pgd_entry =
200 __pgd(_KERNPG_TABLE | __pa(p4d_page_tramp));
201 } else {
202 trampoline_pgd_entry =
203 __pgd(_KERNPG_TABLE | __pa(pud_page_tramp));
204 }
205}