1 .. SPDX-License-Identifier: GPL-2.0 !! 1 The Kernel Address Sanitizer (KASAN)
2 .. Copyright (C) 2023, Google LLC. !! 2 ====================================
3 <<
4 Kernel Address Sanitizer (KASAN) <<
5 ================================ <<
6 3
7 Overview 4 Overview
8 -------- 5 --------
9 6
10 Kernel Address Sanitizer (KASAN) is a dynamic 7 Kernel Address Sanitizer (KASAN) is a dynamic memory safety error detector
11 designed to find out-of-bounds and use-after-f 8 designed to find out-of-bounds and use-after-free bugs.
12 9
13 KASAN has three modes: 10 KASAN has three modes:
14 11
15 1. Generic KASAN 12 1. Generic KASAN
16 2. Software Tag-Based KASAN 13 2. Software Tag-Based KASAN
17 3. Hardware Tag-Based KASAN 14 3. Hardware Tag-Based KASAN
18 15
19 Generic KASAN, enabled with CONFIG_KASAN_GENER 16 Generic KASAN, enabled with CONFIG_KASAN_GENERIC, is the mode intended for
20 debugging, similar to userspace ASan. This mod 17 debugging, similar to userspace ASan. This mode is supported on many CPU
21 architectures, but it has significant performa 18 architectures, but it has significant performance and memory overheads.
22 19
23 Software Tag-Based KASAN or SW_TAGS KASAN, ena 20 Software Tag-Based KASAN or SW_TAGS KASAN, enabled with CONFIG_KASAN_SW_TAGS,
24 can be used for both debugging and dogfood tes 21 can be used for both debugging and dogfood testing, similar to userspace HWASan.
25 This mode is only supported for arm64, but its 22 This mode is only supported for arm64, but its moderate memory overhead allows
26 using it for testing on memory-restricted devi 23 using it for testing on memory-restricted devices with real workloads.
27 24
28 Hardware Tag-Based KASAN or HW_TAGS KASAN, ena 25 Hardware Tag-Based KASAN or HW_TAGS KASAN, enabled with CONFIG_KASAN_HW_TAGS,
29 is the mode intended to be used as an in-field 26 is the mode intended to be used as an in-field memory bug detector or as a
30 security mitigation. This mode only works on a 27 security mitigation. This mode only works on arm64 CPUs that support MTE
31 (Memory Tagging Extension), but it has low mem 28 (Memory Tagging Extension), but it has low memory and performance overheads and
32 thus can be used in production. 29 thus can be used in production.
33 30
34 For details about the memory and performance i 31 For details about the memory and performance impact of each KASAN mode, see the
35 descriptions of the corresponding Kconfig opti 32 descriptions of the corresponding Kconfig options.
36 33
37 The Generic and the Software Tag-Based modes a 34 The Generic and the Software Tag-Based modes are commonly referred to as the
38 software modes. The Software Tag-Based and the 35 software modes. The Software Tag-Based and the Hardware Tag-Based modes are
39 referred to as the tag-based modes. 36 referred to as the tag-based modes.
40 37
41 Support 38 Support
42 ------- 39 -------
43 40
44 Architectures 41 Architectures
45 ~~~~~~~~~~~~~ 42 ~~~~~~~~~~~~~
46 43
47 Generic KASAN is supported on x86_64, arm, arm !! 44 Generic KASAN is supported on x86_64, arm, arm64, powerpc, riscv, s390, and
48 and loongarch, and the tag-based KASAN modes a !! 45 xtensa, and the tag-based KASAN modes are supported only on arm64.
49 46
50 Compilers 47 Compilers
51 ~~~~~~~~~ 48 ~~~~~~~~~
52 49
53 Software KASAN modes use compile-time instrume 50 Software KASAN modes use compile-time instrumentation to insert validity checks
54 before every memory access and thus require a 51 before every memory access and thus require a compiler version that provides
55 support for that. The Hardware Tag-Based mode 52 support for that. The Hardware Tag-Based mode relies on hardware to perform
56 these checks but still requires a compiler ver 53 these checks but still requires a compiler version that supports the memory
57 tagging instructions. 54 tagging instructions.
58 55
59 Generic KASAN requires GCC version 8.3.0 or la 56 Generic KASAN requires GCC version 8.3.0 or later
60 or any Clang version supported by the kernel. 57 or any Clang version supported by the kernel.
61 58
62 Software Tag-Based KASAN requires GCC 11+ 59 Software Tag-Based KASAN requires GCC 11+
63 or any Clang version supported by the kernel. 60 or any Clang version supported by the kernel.
64 61
65 Hardware Tag-Based KASAN requires GCC 10+ or C 62 Hardware Tag-Based KASAN requires GCC 10+ or Clang 12+.
66 63
67 Memory types 64 Memory types
68 ~~~~~~~~~~~~ 65 ~~~~~~~~~~~~
69 66
70 Generic KASAN supports finding bugs in all of 67 Generic KASAN supports finding bugs in all of slab, page_alloc, vmap, vmalloc,
71 stack, and global memory. 68 stack, and global memory.
72 69
73 Software Tag-Based KASAN supports slab, page_a 70 Software Tag-Based KASAN supports slab, page_alloc, vmalloc, and stack memory.
74 71
75 Hardware Tag-Based KASAN supports slab, page_a 72 Hardware Tag-Based KASAN supports slab, page_alloc, and non-executable vmalloc
76 memory. 73 memory.
77 74
78 For slab, both software KASAN modes support SL 75 For slab, both software KASAN modes support SLUB and SLAB allocators, while
79 Hardware Tag-Based KASAN only supports SLUB. 76 Hardware Tag-Based KASAN only supports SLUB.
80 77
81 Usage 78 Usage
82 ----- 79 -----
83 80
84 To enable KASAN, configure the kernel with:: 81 To enable KASAN, configure the kernel with::
85 82
86 CONFIG_KASAN=y 83 CONFIG_KASAN=y
87 84
88 and choose between ``CONFIG_KASAN_GENERIC`` (t 85 and choose between ``CONFIG_KASAN_GENERIC`` (to enable Generic KASAN),
89 ``CONFIG_KASAN_SW_TAGS`` (to enable Software T 86 ``CONFIG_KASAN_SW_TAGS`` (to enable Software Tag-Based KASAN), and
90 ``CONFIG_KASAN_HW_TAGS`` (to enable Hardware T 87 ``CONFIG_KASAN_HW_TAGS`` (to enable Hardware Tag-Based KASAN).
91 88
92 For the software modes, also choose between `` 89 For the software modes, also choose between ``CONFIG_KASAN_OUTLINE`` and
93 ``CONFIG_KASAN_INLINE``. Outline and inline ar 90 ``CONFIG_KASAN_INLINE``. Outline and inline are compiler instrumentation types.
94 The former produces a smaller binary while the 91 The former produces a smaller binary while the latter is up to 2 times faster.
95 92
96 To include alloc and free stack traces of affe 93 To include alloc and free stack traces of affected slab objects into reports,
97 enable ``CONFIG_STACKTRACE``. To include alloc 94 enable ``CONFIG_STACKTRACE``. To include alloc and free stack traces of affected
98 physical pages, enable ``CONFIG_PAGE_OWNER`` a 95 physical pages, enable ``CONFIG_PAGE_OWNER`` and boot with ``page_owner=on``.
99 96
100 Boot parameters 97 Boot parameters
101 ~~~~~~~~~~~~~~~ 98 ~~~~~~~~~~~~~~~
102 99
103 KASAN is affected by the generic ``panic_on_wa 100 KASAN is affected by the generic ``panic_on_warn`` command line parameter.
104 When it is enabled, KASAN panics the kernel af 101 When it is enabled, KASAN panics the kernel after printing a bug report.
105 102
106 By default, KASAN prints a bug report only for 103 By default, KASAN prints a bug report only for the first invalid memory access.
107 With ``kasan_multi_shot``, KASAN prints a repo 104 With ``kasan_multi_shot``, KASAN prints a report on every invalid access. This
108 effectively disables ``panic_on_warn`` for KAS 105 effectively disables ``panic_on_warn`` for KASAN reports.
109 106
110 Alternatively, independent of ``panic_on_warn` 107 Alternatively, independent of ``panic_on_warn``, the ``kasan.fault=`` boot
111 parameter can be used to control panic and rep 108 parameter can be used to control panic and reporting behaviour:
112 109
113 - ``kasan.fault=report``, ``=panic``, or ``=pa !! 110 - ``kasan.fault=report`` or ``=panic`` controls whether to only print a KASAN
114 to only print a KASAN report, panic the kern !! 111 report or also panic the kernel (default: ``report``). The panic happens even
115 invalid writes only (default: ``report``). T !! 112 if ``kasan_multi_shot`` is enabled.
116 ``kasan_multi_shot`` is enabled. Note that w <<
117 Hardware Tag-Based KASAN, ``kasan.fault=pani <<
118 asynchronously checked accesses (including r <<
119 113
120 Software and Hardware Tag-Based KASAN modes (s 114 Software and Hardware Tag-Based KASAN modes (see the section about various
121 modes below) support altering stack trace coll 115 modes below) support altering stack trace collection behavior:
122 116
123 - ``kasan.stacktrace=off`` or ``=on`` disables 117 - ``kasan.stacktrace=off`` or ``=on`` disables or enables alloc and free stack
124 traces collection (default: ``on``). 118 traces collection (default: ``on``).
125 - ``kasan.stack_ring_size=<number of entries>` 119 - ``kasan.stack_ring_size=<number of entries>`` specifies the number of entries
126 in the stack ring (default: ``32768``). 120 in the stack ring (default: ``32768``).
127 121
128 Hardware Tag-Based KASAN mode is intended for 122 Hardware Tag-Based KASAN mode is intended for use in production as a security
129 mitigation. Therefore, it supports additional 123 mitigation. Therefore, it supports additional boot parameters that allow
130 disabling KASAN altogether or controlling its 124 disabling KASAN altogether or controlling its features:
131 125
132 - ``kasan=off`` or ``=on`` controls whether KA 126 - ``kasan=off`` or ``=on`` controls whether KASAN is enabled (default: ``on``).
133 127
134 - ``kasan.mode=sync``, ``=async`` or ``=asymm` 128 - ``kasan.mode=sync``, ``=async`` or ``=asymm`` controls whether KASAN
135 is configured in synchronous, asynchronous o 129 is configured in synchronous, asynchronous or asymmetric mode of
136 execution (default: ``sync``). 130 execution (default: ``sync``).
137 Synchronous mode: a bad access is detected i 131 Synchronous mode: a bad access is detected immediately when a tag
138 check fault occurs. 132 check fault occurs.
139 Asynchronous mode: a bad access detection is 133 Asynchronous mode: a bad access detection is delayed. When a tag check
140 fault occurs, the information is stored in h 134 fault occurs, the information is stored in hardware (in the TFSR_EL1
141 register for arm64). The kernel periodically 135 register for arm64). The kernel periodically checks the hardware and
142 only reports tag faults during these checks. 136 only reports tag faults during these checks.
143 Asymmetric mode: a bad access is detected sy 137 Asymmetric mode: a bad access is detected synchronously on reads and
144 asynchronously on writes. 138 asynchronously on writes.
145 139
146 - ``kasan.vmalloc=off`` or ``=on`` disables or 140 - ``kasan.vmalloc=off`` or ``=on`` disables or enables tagging of vmalloc
147 allocations (default: ``on``). 141 allocations (default: ``on``).
148 142
149 - ``kasan.page_alloc.sample=<sampling interval 143 - ``kasan.page_alloc.sample=<sampling interval>`` makes KASAN tag only every
150 Nth page_alloc allocation with the order equ 144 Nth page_alloc allocation with the order equal or greater than
151 ``kasan.page_alloc.sample.order``, where N i 145 ``kasan.page_alloc.sample.order``, where N is the value of the ``sample``
152 parameter (default: ``1``, or tag every such 146 parameter (default: ``1``, or tag every such allocation).
153 This parameter is intended to mitigate the p 147 This parameter is intended to mitigate the performance overhead introduced
154 by KASAN. 148 by KASAN.
155 Note that enabling this parameter makes Hard 149 Note that enabling this parameter makes Hardware Tag-Based KASAN skip checks
156 of allocations chosen by sampling and thus m 150 of allocations chosen by sampling and thus miss bad accesses to these
157 allocations. Use the default value for accur 151 allocations. Use the default value for accurate bug detection.
158 152
159 - ``kasan.page_alloc.sample.order=<minimum pag 153 - ``kasan.page_alloc.sample.order=<minimum page order>`` specifies the minimum
160 order of allocations that are affected by sa 154 order of allocations that are affected by sampling (default: ``3``).
161 Only applies when ``kasan.page_alloc.sample` 155 Only applies when ``kasan.page_alloc.sample`` is set to a value greater
162 than ``1``. 156 than ``1``.
163 This parameter is intended to allow sampling 157 This parameter is intended to allow sampling only large page_alloc
164 allocations, which is the biggest source of 158 allocations, which is the biggest source of the performance overhead.
165 159
166 Error reports 160 Error reports
167 ~~~~~~~~~~~~~ 161 ~~~~~~~~~~~~~
168 162
169 A typical KASAN report looks like this:: 163 A typical KASAN report looks like this::
170 164
171 ========================================== 165 ==================================================================
172 BUG: KASAN: slab-out-of-bounds in kmalloc_ !! 166 BUG: KASAN: slab-out-of-bounds in kmalloc_oob_right+0xa8/0xbc [test_kasan]
173 Write of size 1 at addr ffff8801f44ec37b b 167 Write of size 1 at addr ffff8801f44ec37b by task insmod/2760
174 168
175 CPU: 1 PID: 2760 Comm: insmod Not tainted 169 CPU: 1 PID: 2760 Comm: insmod Not tainted 4.19.0-rc3+ #698
176 Hardware name: QEMU Standard PC (i440FX + 170 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014
177 Call Trace: 171 Call Trace:
178 dump_stack+0x94/0xd8 172 dump_stack+0x94/0xd8
179 print_address_description+0x73/0x280 173 print_address_description+0x73/0x280
180 kasan_report+0x144/0x187 174 kasan_report+0x144/0x187
181 __asan_report_store1_noabort+0x17/0x20 175 __asan_report_store1_noabort+0x17/0x20
182 kmalloc_oob_right+0xa8/0xbc [kasan_test] !! 176 kmalloc_oob_right+0xa8/0xbc [test_kasan]
183 kmalloc_tests_init+0x16/0x700 [kasan_test !! 177 kmalloc_tests_init+0x16/0x700 [test_kasan]
184 do_one_initcall+0xa5/0x3ae 178 do_one_initcall+0xa5/0x3ae
185 do_init_module+0x1b6/0x547 179 do_init_module+0x1b6/0x547
186 load_module+0x75df/0x8070 180 load_module+0x75df/0x8070
187 __do_sys_init_module+0x1c6/0x200 181 __do_sys_init_module+0x1c6/0x200
188 __x64_sys_init_module+0x6e/0xb0 182 __x64_sys_init_module+0x6e/0xb0
189 do_syscall_64+0x9f/0x2c0 183 do_syscall_64+0x9f/0x2c0
190 entry_SYSCALL_64_after_hwframe+0x44/0xa9 184 entry_SYSCALL_64_after_hwframe+0x44/0xa9
191 RIP: 0033:0x7f96443109da 185 RIP: 0033:0x7f96443109da
192 RSP: 002b:00007ffcf0b51b08 EFLAGS: 0000020 186 RSP: 002b:00007ffcf0b51b08 EFLAGS: 00000202 ORIG_RAX: 00000000000000af
193 RAX: ffffffffffffffda RBX: 000055dc3ee521a 187 RAX: ffffffffffffffda RBX: 000055dc3ee521a0 RCX: 00007f96443109da
194 RDX: 00007f96445cff88 RSI: 0000000000057a5 188 RDX: 00007f96445cff88 RSI: 0000000000057a50 RDI: 00007f9644992000
195 RBP: 000055dc3ee510b0 R08: 000000000000000 189 RBP: 000055dc3ee510b0 R08: 0000000000000003 R09: 0000000000000000
196 R10: 00007f964430cd0a R11: 000000000000020 190 R10: 00007f964430cd0a R11: 0000000000000202 R12: 00007f96445cff88
197 R13: 000055dc3ee51090 R14: 000000000000000 191 R13: 000055dc3ee51090 R14: 0000000000000000 R15: 0000000000000000
198 192
199 Allocated by task 2760: 193 Allocated by task 2760:
200 save_stack+0x43/0xd0 194 save_stack+0x43/0xd0
201 kasan_kmalloc+0xa7/0xd0 195 kasan_kmalloc+0xa7/0xd0
202 kmem_cache_alloc_trace+0xe1/0x1b0 196 kmem_cache_alloc_trace+0xe1/0x1b0
203 kmalloc_oob_right+0x56/0xbc [kasan_test] !! 197 kmalloc_oob_right+0x56/0xbc [test_kasan]
204 kmalloc_tests_init+0x16/0x700 [kasan_test !! 198 kmalloc_tests_init+0x16/0x700 [test_kasan]
205 do_one_initcall+0xa5/0x3ae 199 do_one_initcall+0xa5/0x3ae
206 do_init_module+0x1b6/0x547 200 do_init_module+0x1b6/0x547
207 load_module+0x75df/0x8070 201 load_module+0x75df/0x8070
208 __do_sys_init_module+0x1c6/0x200 202 __do_sys_init_module+0x1c6/0x200
209 __x64_sys_init_module+0x6e/0xb0 203 __x64_sys_init_module+0x6e/0xb0
210 do_syscall_64+0x9f/0x2c0 204 do_syscall_64+0x9f/0x2c0
211 entry_SYSCALL_64_after_hwframe+0x44/0xa9 205 entry_SYSCALL_64_after_hwframe+0x44/0xa9
212 206
213 Freed by task 815: 207 Freed by task 815:
214 save_stack+0x43/0xd0 208 save_stack+0x43/0xd0
215 __kasan_slab_free+0x135/0x190 209 __kasan_slab_free+0x135/0x190
216 kasan_slab_free+0xe/0x10 210 kasan_slab_free+0xe/0x10
217 kfree+0x93/0x1a0 211 kfree+0x93/0x1a0
218 umh_complete+0x6a/0xa0 212 umh_complete+0x6a/0xa0
219 call_usermodehelper_exec_async+0x4c3/0x64 213 call_usermodehelper_exec_async+0x4c3/0x640
220 ret_from_fork+0x35/0x40 214 ret_from_fork+0x35/0x40
221 215
222 The buggy address belongs to the object at 216 The buggy address belongs to the object at ffff8801f44ec300
223 which belongs to the cache kmalloc-128 of 217 which belongs to the cache kmalloc-128 of size 128
224 The buggy address is located 123 bytes ins 218 The buggy address is located 123 bytes inside of
225 128-byte region [ffff8801f44ec300, ffff88 219 128-byte region [ffff8801f44ec300, ffff8801f44ec380)
226 The buggy address belongs to the page: 220 The buggy address belongs to the page:
227 page:ffffea0007d13b00 count:1 mapcount:0 m 221 page:ffffea0007d13b00 count:1 mapcount:0 mapping:ffff8801f7001640 index:0x0
228 flags: 0x200000000000100(slab) 222 flags: 0x200000000000100(slab)
229 raw: 0200000000000100 ffffea0007d11dc0 000 223 raw: 0200000000000100 ffffea0007d11dc0 0000001a0000001a ffff8801f7001640
230 raw: 0000000000000000 0000000080150015 000 224 raw: 0000000000000000 0000000080150015 00000001ffffffff 0000000000000000
231 page dumped because: kasan: bad access det 225 page dumped because: kasan: bad access detected
232 226
233 Memory state around the buggy address: 227 Memory state around the buggy address:
234 ffff8801f44ec200: fc fc fc fc fc fc fc fc 228 ffff8801f44ec200: fc fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb
235 ffff8801f44ec280: fb fb fb fb fb fb fb fb 229 ffff8801f44ec280: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
236 >ffff8801f44ec300: 00 00 00 00 00 00 00 00 230 >ffff8801f44ec300: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 03
237 231 ^
238 ffff8801f44ec380: fc fc fc fc fc fc fc fc 232 ffff8801f44ec380: fc fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb
239 ffff8801f44ec400: fb fb fb fb fb fb fb fb 233 ffff8801f44ec400: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
240 ========================================== 234 ==================================================================
241 235
242 The report header summarizes what kind of bug 236 The report header summarizes what kind of bug happened and what kind of access
243 caused it. It is followed by a stack trace of 237 caused it. It is followed by a stack trace of the bad access, a stack trace of
244 where the accessed memory was allocated (in ca 238 where the accessed memory was allocated (in case a slab object was accessed),
245 and a stack trace of where the object was free 239 and a stack trace of where the object was freed (in case of a use-after-free
246 bug report). Next comes a description of the a 240 bug report). Next comes a description of the accessed slab object and the
247 information about the accessed memory page. 241 information about the accessed memory page.
248 242
249 In the end, the report shows the memory state 243 In the end, the report shows the memory state around the accessed address.
250 Internally, KASAN tracks memory state separate 244 Internally, KASAN tracks memory state separately for each memory granule, which
251 is either 8 or 16 aligned bytes depending on K 245 is either 8 or 16 aligned bytes depending on KASAN mode. Each number in the
252 memory state section of the report shows the s 246 memory state section of the report shows the state of one of the memory
253 granules that surround the accessed address. 247 granules that surround the accessed address.
254 248
255 For Generic KASAN, the size of each memory gra 249 For Generic KASAN, the size of each memory granule is 8. The state of each
256 granule is encoded in one shadow byte. Those 8 250 granule is encoded in one shadow byte. Those 8 bytes can be accessible,
257 partially accessible, freed, or be a part of a 251 partially accessible, freed, or be a part of a redzone. KASAN uses the following
258 encoding for each shadow byte: 00 means that a 252 encoding for each shadow byte: 00 means that all 8 bytes of the corresponding
259 memory region are accessible; number N (1 <= N 253 memory region are accessible; number N (1 <= N <= 7) means that the first N
260 bytes are accessible, and other (8 - N) bytes 254 bytes are accessible, and other (8 - N) bytes are not; any negative value
261 indicates that the entire 8-byte word is inacc 255 indicates that the entire 8-byte word is inaccessible. KASAN uses different
262 negative values to distinguish between differe 256 negative values to distinguish between different kinds of inaccessible memory
263 like redzones or freed memory (see mm/kasan/ka 257 like redzones or freed memory (see mm/kasan/kasan.h).
264 258
265 In the report above, the arrow points to the s 259 In the report above, the arrow points to the shadow byte ``03``, which means
266 that the accessed address is partially accessi 260 that the accessed address is partially accessible.
267 261
268 For tag-based KASAN modes, this last report se 262 For tag-based KASAN modes, this last report section shows the memory tags around
269 the accessed address (see the `Implementation 263 the accessed address (see the `Implementation details`_ section).
270 264
271 Note that KASAN bug titles (like ``slab-out-of 265 Note that KASAN bug titles (like ``slab-out-of-bounds`` or ``use-after-free``)
272 are best-effort: KASAN prints the most probabl 266 are best-effort: KASAN prints the most probable bug type based on the limited
273 information it has. The actual type of the bug 267 information it has. The actual type of the bug might be different.
274 268
275 Generic KASAN also reports up to two auxiliary 269 Generic KASAN also reports up to two auxiliary call stack traces. These stack
276 traces point to places in code that interacted 270 traces point to places in code that interacted with the object but that are not
277 directly present in the bad access stack trace 271 directly present in the bad access stack trace. Currently, this includes
278 call_rcu() and workqueue queuing. 272 call_rcu() and workqueue queuing.
279 273
280 CONFIG_KASAN_EXTRA_INFO <<
281 ~~~~~~~~~~~~~~~~~~~~~~~ <<
282 <<
283 Enabling CONFIG_KASAN_EXTRA_INFO allows KASAN <<
284 information. The extra information currently s <<
285 timestamp at allocation and free. More informa <<
286 the bug and correlate the error with other sys <<
287 extra memory to record more information (more <<
288 CONFIG_KASAN_EXTRA_INFO). <<
289 <<
290 Here is the report with CONFIG_KASAN_EXTRA_INF <<
291 different parts are shown):: <<
292 <<
293 ========================================== <<
294 ... <<
295 Allocated by task 134 on cpu 5 at 229.1338 <<
296 ... <<
297 Freed by task 136 on cpu 3 at 230.199335s: <<
298 ... <<
299 ========================================== <<
300 <<
301 Implementation details 274 Implementation details
302 ---------------------- 275 ----------------------
303 276
304 Generic KASAN 277 Generic KASAN
305 ~~~~~~~~~~~~~ 278 ~~~~~~~~~~~~~
306 279
307 Software KASAN modes use shadow memory to reco 280 Software KASAN modes use shadow memory to record whether each byte of memory is
308 safe to access and use compile-time instrument 281 safe to access and use compile-time instrumentation to insert shadow memory
309 checks before each memory access. 282 checks before each memory access.
310 283
311 Generic KASAN dedicates 1/8th of kernel memory 284 Generic KASAN dedicates 1/8th of kernel memory to its shadow memory (16TB
312 to cover 128TB on x86_64) and uses direct mapp 285 to cover 128TB on x86_64) and uses direct mapping with a scale and offset to
313 translate a memory address to its correspondin 286 translate a memory address to its corresponding shadow address.
314 287
315 Here is the function which translates an addre 288 Here is the function which translates an address to its corresponding shadow
316 address:: 289 address::
317 290
318 static inline void *kasan_mem_to_shadow(co 291 static inline void *kasan_mem_to_shadow(const void *addr)
319 { 292 {
320 return (void *)((unsigned long)addr >> 293 return (void *)((unsigned long)addr >> KASAN_SHADOW_SCALE_SHIFT)
321 + KASAN_SHADOW_OFFSET; 294 + KASAN_SHADOW_OFFSET;
322 } 295 }
323 296
324 where ``KASAN_SHADOW_SCALE_SHIFT = 3``. 297 where ``KASAN_SHADOW_SCALE_SHIFT = 3``.
325 298
326 Compile-time instrumentation is used to insert 299 Compile-time instrumentation is used to insert memory access checks. Compiler
327 inserts function calls (``__asan_load*(addr)`` 300 inserts function calls (``__asan_load*(addr)``, ``__asan_store*(addr)``) before
328 each memory access of size 1, 2, 4, 8, or 16. 301 each memory access of size 1, 2, 4, 8, or 16. These functions check whether
329 memory accesses are valid or not by checking c 302 memory accesses are valid or not by checking corresponding shadow memory.
330 303
331 With inline instrumentation, instead of making 304 With inline instrumentation, instead of making function calls, the compiler
332 directly inserts the code to check shadow memo 305 directly inserts the code to check shadow memory. This option significantly
333 enlarges the kernel, but it gives an x1.1-x2 p 306 enlarges the kernel, but it gives an x1.1-x2 performance boost over the
334 outline-instrumented kernel. 307 outline-instrumented kernel.
335 308
336 Generic KASAN is the only mode that delays the 309 Generic KASAN is the only mode that delays the reuse of freed objects via
337 quarantine (see mm/kasan/quarantine.c for impl 310 quarantine (see mm/kasan/quarantine.c for implementation).
338 311
339 Software Tag-Based KASAN 312 Software Tag-Based KASAN
340 ~~~~~~~~~~~~~~~~~~~~~~~~ 313 ~~~~~~~~~~~~~~~~~~~~~~~~
341 314
342 Software Tag-Based KASAN uses a software memor 315 Software Tag-Based KASAN uses a software memory tagging approach to checking
343 access validity. It is currently only implemen 316 access validity. It is currently only implemented for the arm64 architecture.
344 317
345 Software Tag-Based KASAN uses the Top Byte Ign 318 Software Tag-Based KASAN uses the Top Byte Ignore (TBI) feature of arm64 CPUs
346 to store a pointer tag in the top byte of kern 319 to store a pointer tag in the top byte of kernel pointers. It uses shadow memory
347 to store memory tags associated with each 16-b 320 to store memory tags associated with each 16-byte memory cell (therefore, it
348 dedicates 1/16th of the kernel memory for shad 321 dedicates 1/16th of the kernel memory for shadow memory).
349 322
350 On each memory allocation, Software Tag-Based 323 On each memory allocation, Software Tag-Based KASAN generates a random tag, tags
351 the allocated memory with this tag, and embeds 324 the allocated memory with this tag, and embeds the same tag into the returned
352 pointer. 325 pointer.
353 326
354 Software Tag-Based KASAN uses compile-time ins 327 Software Tag-Based KASAN uses compile-time instrumentation to insert checks
355 before each memory access. These checks make s 328 before each memory access. These checks make sure that the tag of the memory
356 that is being accessed is equal to the tag of 329 that is being accessed is equal to the tag of the pointer that is used to access
357 this memory. In case of a tag mismatch, Softwa 330 this memory. In case of a tag mismatch, Software Tag-Based KASAN prints a bug
358 report. 331 report.
359 332
360 Software Tag-Based KASAN also has two instrume 333 Software Tag-Based KASAN also has two instrumentation modes (outline, which
361 emits callbacks to check memory accesses; and 334 emits callbacks to check memory accesses; and inline, which performs the shadow
362 memory checks inline). With outline instrument 335 memory checks inline). With outline instrumentation mode, a bug report is
363 printed from the function that performs the ac 336 printed from the function that performs the access check. With inline
364 instrumentation, a ``brk`` instruction is emit 337 instrumentation, a ``brk`` instruction is emitted by the compiler, and a
365 dedicated ``brk`` handler is used to print bug 338 dedicated ``brk`` handler is used to print bug reports.
366 339
367 Software Tag-Based KASAN uses 0xFF as a match- 340 Software Tag-Based KASAN uses 0xFF as a match-all pointer tag (accesses through
368 pointers with the 0xFF pointer tag are not che 341 pointers with the 0xFF pointer tag are not checked). The value 0xFE is currently
369 reserved to tag freed memory regions. 342 reserved to tag freed memory regions.
370 343
371 Hardware Tag-Based KASAN 344 Hardware Tag-Based KASAN
372 ~~~~~~~~~~~~~~~~~~~~~~~~ 345 ~~~~~~~~~~~~~~~~~~~~~~~~
373 346
374 Hardware Tag-Based KASAN is similar to the sof 347 Hardware Tag-Based KASAN is similar to the software mode in concept but uses
375 hardware memory tagging support instead of com 348 hardware memory tagging support instead of compiler instrumentation and
376 shadow memory. 349 shadow memory.
377 350
378 Hardware Tag-Based KASAN is currently only imp 351 Hardware Tag-Based KASAN is currently only implemented for arm64 architecture
379 and based on both arm64 Memory Tagging Extensi 352 and based on both arm64 Memory Tagging Extension (MTE) introduced in ARMv8.5
380 Instruction Set Architecture and Top Byte Igno 353 Instruction Set Architecture and Top Byte Ignore (TBI).
381 354
382 Special arm64 instructions are used to assign 355 Special arm64 instructions are used to assign memory tags for each allocation.
383 Same tags are assigned to pointers to those al 356 Same tags are assigned to pointers to those allocations. On every memory
384 access, hardware makes sure that the tag of th 357 access, hardware makes sure that the tag of the memory that is being accessed is
385 equal to the tag of the pointer that is used t 358 equal to the tag of the pointer that is used to access this memory. In case of a
386 tag mismatch, a fault is generated, and a repo 359 tag mismatch, a fault is generated, and a report is printed.
387 360
388 Hardware Tag-Based KASAN uses 0xFF as a match- 361 Hardware Tag-Based KASAN uses 0xFF as a match-all pointer tag (accesses through
389 pointers with the 0xFF pointer tag are not che 362 pointers with the 0xFF pointer tag are not checked). The value 0xFE is currently
390 reserved to tag freed memory regions. 363 reserved to tag freed memory regions.
391 364
392 If the hardware does not support MTE (pre ARMv 365 If the hardware does not support MTE (pre ARMv8.5), Hardware Tag-Based KASAN
393 will not be enabled. In this case, all KASAN b 366 will not be enabled. In this case, all KASAN boot parameters are ignored.
394 367
395 Note that enabling CONFIG_KASAN_HW_TAGS always 368 Note that enabling CONFIG_KASAN_HW_TAGS always results in in-kernel TBI being
396 enabled. Even when ``kasan.mode=off`` is provi 369 enabled. Even when ``kasan.mode=off`` is provided or when the hardware does not
397 support MTE (but supports TBI). 370 support MTE (but supports TBI).
398 371
399 Hardware Tag-Based KASAN only reports the firs 372 Hardware Tag-Based KASAN only reports the first found bug. After that, MTE tag
400 checking gets disabled. 373 checking gets disabled.
401 374
402 Shadow memory 375 Shadow memory
403 ------------- 376 -------------
404 377
405 The contents of this section are only applicab 378 The contents of this section are only applicable to software KASAN modes.
406 379
407 The kernel maps memory in several different pa 380 The kernel maps memory in several different parts of the address space.
408 The range of kernel virtual addresses is large 381 The range of kernel virtual addresses is large: there is not enough real
409 memory to support a real shadow region for eve 382 memory to support a real shadow region for every address that could be
410 accessed by the kernel. Therefore, KASAN only 383 accessed by the kernel. Therefore, KASAN only maps real shadow for certain
411 parts of the address space. 384 parts of the address space.
412 385
413 Default behaviour 386 Default behaviour
414 ~~~~~~~~~~~~~~~~~ 387 ~~~~~~~~~~~~~~~~~
415 388
416 By default, architectures only map real memory 389 By default, architectures only map real memory over the shadow region
417 for the linear mapping (and potentially other 390 for the linear mapping (and potentially other small areas). For all
418 other areas - such as vmalloc and vmemmap spac 391 other areas - such as vmalloc and vmemmap space - a single read-only
419 page is mapped over the shadow area. This read 392 page is mapped over the shadow area. This read-only shadow page
420 declares all memory accesses as permitted. 393 declares all memory accesses as permitted.
421 394
422 This presents a problem for modules: they do n 395 This presents a problem for modules: they do not live in the linear
423 mapping but in a dedicated module space. By ho 396 mapping but in a dedicated module space. By hooking into the module
424 allocator, KASAN temporarily maps real shadow 397 allocator, KASAN temporarily maps real shadow memory to cover them.
425 This allows detection of invalid accesses to m 398 This allows detection of invalid accesses to module globals, for example.
426 399
427 This also creates an incompatibility with ``VM 400 This also creates an incompatibility with ``VMAP_STACK``: if the stack
428 lives in vmalloc space, it will be shadowed by 401 lives in vmalloc space, it will be shadowed by the read-only page, and
429 the kernel will fault when trying to set up th 402 the kernel will fault when trying to set up the shadow data for stack
430 variables. 403 variables.
431 404
432 CONFIG_KASAN_VMALLOC 405 CONFIG_KASAN_VMALLOC
433 ~~~~~~~~~~~~~~~~~~~~ 406 ~~~~~~~~~~~~~~~~~~~~
434 407
435 With ``CONFIG_KASAN_VMALLOC``, KASAN can cover 408 With ``CONFIG_KASAN_VMALLOC``, KASAN can cover vmalloc space at the
436 cost of greater memory usage. Currently, this 409 cost of greater memory usage. Currently, this is supported on x86,
437 arm64, riscv, s390, and powerpc. 410 arm64, riscv, s390, and powerpc.
438 411
439 This works by hooking into vmalloc and vmap an 412 This works by hooking into vmalloc and vmap and dynamically
440 allocating real shadow memory to back the mapp 413 allocating real shadow memory to back the mappings.
441 414
442 Most mappings in vmalloc space are small, requ 415 Most mappings in vmalloc space are small, requiring less than a full
443 page of shadow space. Allocating a full shadow 416 page of shadow space. Allocating a full shadow page per mapping would
444 therefore be wasteful. Furthermore, to ensure 417 therefore be wasteful. Furthermore, to ensure that different mappings
445 use different shadow pages, mappings would hav 418 use different shadow pages, mappings would have to be aligned to
446 ``KASAN_GRANULE_SIZE * PAGE_SIZE``. 419 ``KASAN_GRANULE_SIZE * PAGE_SIZE``.
447 420
448 Instead, KASAN shares backing space across mul 421 Instead, KASAN shares backing space across multiple mappings. It allocates
449 a backing page when a mapping in vmalloc space 422 a backing page when a mapping in vmalloc space uses a particular page
450 of the shadow region. This page can be shared 423 of the shadow region. This page can be shared by other vmalloc
451 mappings later on. 424 mappings later on.
452 425
453 KASAN hooks into the vmap infrastructure to la 426 KASAN hooks into the vmap infrastructure to lazily clean up unused shadow
454 memory. 427 memory.
455 428
456 To avoid the difficulties around swapping mapp 429 To avoid the difficulties around swapping mappings around, KASAN expects
457 that the part of the shadow region that covers 430 that the part of the shadow region that covers the vmalloc space will
458 not be covered by the early shadow page but wi 431 not be covered by the early shadow page but will be left unmapped.
459 This will require changes in arch-specific cod 432 This will require changes in arch-specific code.
460 433
461 This allows ``VMAP_STACK`` support on x86 and 434 This allows ``VMAP_STACK`` support on x86 and can simplify support of
462 architectures that do not have a fixed module 435 architectures that do not have a fixed module region.
463 436
464 For developers 437 For developers
465 -------------- 438 --------------
466 439
467 Ignoring accesses 440 Ignoring accesses
468 ~~~~~~~~~~~~~~~~~ 441 ~~~~~~~~~~~~~~~~~
469 442
470 Software KASAN modes use compiler instrumentat 443 Software KASAN modes use compiler instrumentation to insert validity checks.
471 Such instrumentation might be incompatible wit 444 Such instrumentation might be incompatible with some parts of the kernel, and
472 therefore needs to be disabled. 445 therefore needs to be disabled.
473 446
474 Other parts of the kernel might access metadat 447 Other parts of the kernel might access metadata for allocated objects.
475 Normally, KASAN detects and reports such acces 448 Normally, KASAN detects and reports such accesses, but in some cases (e.g.,
476 in memory allocators), these accesses are vali 449 in memory allocators), these accesses are valid.
477 450
478 For software KASAN modes, to disable instrumen 451 For software KASAN modes, to disable instrumentation for a specific file or
479 directory, add a ``KASAN_SANITIZE`` annotation 452 directory, add a ``KASAN_SANITIZE`` annotation to the respective kernel
480 Makefile: 453 Makefile:
481 454
482 - For a single file (e.g., main.o):: 455 - For a single file (e.g., main.o)::
483 456
484 KASAN_SANITIZE_main.o := n 457 KASAN_SANITIZE_main.o := n
485 458
486 - For all files in one directory:: 459 - For all files in one directory::
487 460
488 KASAN_SANITIZE := n 461 KASAN_SANITIZE := n
489 462
490 For software KASAN modes, to disable instrumen 463 For software KASAN modes, to disable instrumentation on a per-function basis,
491 use the KASAN-specific ``__no_sanitize_address 464 use the KASAN-specific ``__no_sanitize_address`` function attribute or the
492 generic ``noinstr`` one. 465 generic ``noinstr`` one.
493 466
494 Note that disabling compiler instrumentation ( 467 Note that disabling compiler instrumentation (either on a per-file or a
495 per-function basis) makes KASAN ignore the acc 468 per-function basis) makes KASAN ignore the accesses that happen directly in
496 that code for software KASAN modes. It does no 469 that code for software KASAN modes. It does not help when the accesses happen
497 indirectly (through calls to instrumented func 470 indirectly (through calls to instrumented functions) or with Hardware
498 Tag-Based KASAN, which does not use compiler i 471 Tag-Based KASAN, which does not use compiler instrumentation.
499 472
500 For software KASAN modes, to disable KASAN rep 473 For software KASAN modes, to disable KASAN reports in a part of the kernel code
501 for the current task, annotate this part of th 474 for the current task, annotate this part of the code with a
502 ``kasan_disable_current()``/``kasan_enable_cur 475 ``kasan_disable_current()``/``kasan_enable_current()`` section. This also
503 disables the reports for indirect accesses tha 476 disables the reports for indirect accesses that happen through function calls.
504 477
505 For tag-based KASAN modes, to disable access c 478 For tag-based KASAN modes, to disable access checking, use
506 ``kasan_reset_tag()`` or ``page_kasan_tag_rese 479 ``kasan_reset_tag()`` or ``page_kasan_tag_reset()``. Note that temporarily
507 disabling access checking via ``page_kasan_tag 480 disabling access checking via ``page_kasan_tag_reset()`` requires saving and
508 restoring the per-page KASAN tag via ``page_ka 481 restoring the per-page KASAN tag via ``page_kasan_tag``/``page_kasan_tag_set``.
509 482
510 Tests 483 Tests
511 ~~~~~ 484 ~~~~~
512 485
513 There are KASAN tests that allow verifying tha 486 There are KASAN tests that allow verifying that KASAN works and can detect
514 certain types of memory corruptions. The tests 487 certain types of memory corruptions. The tests consist of two parts:
515 488
516 1. Tests that are integrated with the KUnit Te 489 1. Tests that are integrated with the KUnit Test Framework. Enabled with
517 ``CONFIG_KASAN_KUNIT_TEST``. These tests can b 490 ``CONFIG_KASAN_KUNIT_TEST``. These tests can be run and partially verified
518 automatically in a few different ways; see the 491 automatically in a few different ways; see the instructions below.
519 492
520 2. Tests that are currently incompatible with 493 2. Tests that are currently incompatible with KUnit. Enabled with
521 ``CONFIG_KASAN_MODULE_TEST`` and can only be r 494 ``CONFIG_KASAN_MODULE_TEST`` and can only be run as a module. These tests can
522 only be verified manually by loading the kerne 495 only be verified manually by loading the kernel module and inspecting the
523 kernel log for KASAN reports. 496 kernel log for KASAN reports.
524 497
525 Each KUnit-compatible KASAN test prints one of 498 Each KUnit-compatible KASAN test prints one of multiple KASAN reports if an
526 error is detected. Then the test prints its nu 499 error is detected. Then the test prints its number and status.
527 500
528 When a test passes:: 501 When a test passes::
529 502
530 ok 28 - kmalloc_double_kzfree 503 ok 28 - kmalloc_double_kzfree
531 504
532 When a test fails due to a failed ``kmalloc``: 505 When a test fails due to a failed ``kmalloc``::
533 506
534 # kmalloc_large_oob_right: ASSERTION F !! 507 # kmalloc_large_oob_right: ASSERTION FAILED at lib/test_kasan.c:163
535 Expected ptr is not null, but is 508 Expected ptr is not null, but is
536 not ok 5 - kmalloc_large_oob_right !! 509 not ok 4 - kmalloc_large_oob_right
537 510
538 When a test fails due to a missing KASAN repor 511 When a test fails due to a missing KASAN report::
539 512
540 # kmalloc_double_kzfree: EXPECTATION F !! 513 # kmalloc_double_kzfree: EXPECTATION FAILED at lib/test_kasan.c:974
541 KASAN failure expected in "kfree_sensi 514 KASAN failure expected in "kfree_sensitive(ptr)", but none occurred
542 not ok 28 - kmalloc_double_kzfree !! 515 not ok 44 - kmalloc_double_kzfree
543 516
544 517
545 At the end the cumulative status of all KASAN 518 At the end the cumulative status of all KASAN tests is printed. On success::
546 519
547 ok 1 - kasan 520 ok 1 - kasan
548 521
549 Or, if one of the tests failed:: 522 Or, if one of the tests failed::
550 523
551 not ok 1 - kasan 524 not ok 1 - kasan
552 525
553 There are a few ways to run KUnit-compatible K 526 There are a few ways to run KUnit-compatible KASAN tests.
554 527
555 1. Loadable module 528 1. Loadable module
556 529
557 With ``CONFIG_KUNIT`` enabled, KASAN-KUnit 530 With ``CONFIG_KUNIT`` enabled, KASAN-KUnit tests can be built as a loadable
558 module and run by loading ``kasan_test.ko`` !! 531 module and run by loading ``test_kasan.ko`` with ``insmod`` or ``modprobe``.
559 532
560 2. Built-In 533 2. Built-In
561 534
562 With ``CONFIG_KUNIT`` built-in, KASAN-KUnit 535 With ``CONFIG_KUNIT`` built-in, KASAN-KUnit tests can be built-in as well.
563 In this case, the tests will run at boot as 536 In this case, the tests will run at boot as a late-init call.
564 537
565 3. Using kunit_tool 538 3. Using kunit_tool
566 539
567 With ``CONFIG_KUNIT`` and ``CONFIG_KASAN_KU 540 With ``CONFIG_KUNIT`` and ``CONFIG_KASAN_KUNIT_TEST`` built-in, it is also
568 possible to use ``kunit_tool`` to see the r 541 possible to use ``kunit_tool`` to see the results of KUnit tests in a more
569 readable way. This will not print the KASAN 542 readable way. This will not print the KASAN reports of the tests that passed.
570 See `KUnit documentation <https://www.kerne 543 See `KUnit documentation <https://www.kernel.org/doc/html/latest/dev-tools/kunit/index.html>`_
571 for more up-to-date information on ``kunit_ 544 for more up-to-date information on ``kunit_tool``.
572 545
573 .. _KUnit: https://www.kernel.org/doc/html/lat 546 .. _KUnit: https://www.kernel.org/doc/html/latest/dev-tools/kunit/index.html
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.