1 The Kernel Address Sanitizer (KASAN)
2 ====================================
7 KernelAddressSANitizer (KASAN) is a dynamic memory error detector designed to
8 find out-of-bound and use-after-free bugs. KASAN has two modes: generic KASAN
9 (similar to userspace ASan) and software tag-based KASAN (similar to userspace
12 KASAN uses compile-time instrumentation to insert validity checks before every
13 memory access, and therefore requires a compiler version that supports that.
15 Generic KASAN is supported in both GCC and Clang. With GCC it requires version
16 8.3.0 or later. With Clang it requires version 7.0.0 or later, but detection of
17 out-of-bounds accesses for global variables is only supported since Clang 11.
19 Tag-based KASAN is only supported in Clang and requires version 7.0.0 or later.
21 Currently generic KASAN is supported for the x86_64, arm64, xtensa, s390 and
22 riscv architectures, and tag-based KASAN is supported only for arm64.
27 To enable KASAN configure kernel with::
31 and choose between CONFIG_KASAN_GENERIC (to enable generic KASAN) and
32 CONFIG_KASAN_SW_TAGS (to enable software tag-based KASAN).
34 You also need to choose between CONFIG_KASAN_OUTLINE and CONFIG_KASAN_INLINE.
35 Outline and inline are compiler instrumentation types. The former produces
36 smaller binary while the latter is 1.1 - 2 times faster.
38 Both KASAN modes work with both SLUB and SLAB memory allocators.
39 For better bug detection and nicer reporting, enable CONFIG_STACKTRACE.
41 To augment reports with last allocation and freeing stack of the physical page,
42 it is recommended to enable also CONFIG_PAGE_OWNER and boot with page_owner=on.
44 To disable instrumentation for specific files or directories, add a line
45 similar to the following to the respective kernel Makefile:
47 - For a single file (e.g. main.o)::
49 KASAN_SANITIZE_main.o := n
51 - For all files in one directory::
58 A typical out-of-bounds access generic KASAN report looks like this::
60 ==================================================================
61 BUG: KASAN: slab-out-of-bounds in kmalloc_oob_right+0xa8/0xbc [test_kasan]
62 Write of size 1 at addr ffff8801f44ec37b by task insmod/2760
64 CPU: 1 PID: 2760 Comm: insmod Not tainted 4.19.0-rc3+ #698
65 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1 04/01/2014
68 print_address_description+0x73/0x280
69 kasan_report+0x144/0x187
70 __asan_report_store1_noabort+0x17/0x20
71 kmalloc_oob_right+0xa8/0xbc [test_kasan]
72 kmalloc_tests_init+0x16/0x700 [test_kasan]
73 do_one_initcall+0xa5/0x3ae
74 do_init_module+0x1b6/0x547
75 load_module+0x75df/0x8070
76 __do_sys_init_module+0x1c6/0x200
77 __x64_sys_init_module+0x6e/0xb0
78 do_syscall_64+0x9f/0x2c0
79 entry_SYSCALL_64_after_hwframe+0x44/0xa9
80 RIP: 0033:0x7f96443109da
81 RSP: 002b:00007ffcf0b51b08 EFLAGS: 00000202 ORIG_RAX: 00000000000000af
82 RAX: ffffffffffffffda RBX: 000055dc3ee521a0 RCX: 00007f96443109da
83 RDX: 00007f96445cff88 RSI: 0000000000057a50 RDI: 00007f9644992000
84 RBP: 000055dc3ee510b0 R08: 0000000000000003 R09: 0000000000000000
85 R10: 00007f964430cd0a R11: 0000000000000202 R12: 00007f96445cff88
86 R13: 000055dc3ee51090 R14: 0000000000000000 R15: 0000000000000000
88 Allocated by task 2760:
90 kasan_kmalloc+0xa7/0xd0
91 kmem_cache_alloc_trace+0xe1/0x1b0
92 kmalloc_oob_right+0x56/0xbc [test_kasan]
93 kmalloc_tests_init+0x16/0x700 [test_kasan]
94 do_one_initcall+0xa5/0x3ae
95 do_init_module+0x1b6/0x547
96 load_module+0x75df/0x8070
97 __do_sys_init_module+0x1c6/0x200
98 __x64_sys_init_module+0x6e/0xb0
99 do_syscall_64+0x9f/0x2c0
100 entry_SYSCALL_64_after_hwframe+0x44/0xa9
104 __kasan_slab_free+0x135/0x190
105 kasan_slab_free+0xe/0x10
107 umh_complete+0x6a/0xa0
108 call_usermodehelper_exec_async+0x4c3/0x640
109 ret_from_fork+0x35/0x40
111 The buggy address belongs to the object at ffff8801f44ec300
112 which belongs to the cache kmalloc-128 of size 128
113 The buggy address is located 123 bytes inside of
114 128-byte region [ffff8801f44ec300, ffff8801f44ec380)
115 The buggy address belongs to the page:
116 page:ffffea0007d13b00 count:1 mapcount:0 mapping:ffff8801f7001640 index:0x0
117 flags: 0x200000000000100(slab)
118 raw: 0200000000000100 ffffea0007d11dc0 0000001a0000001a ffff8801f7001640
119 raw: 0000000000000000 0000000080150015 00000001ffffffff 0000000000000000
120 page dumped because: kasan: bad access detected
122 Memory state around the buggy address:
123 ffff8801f44ec200: fc fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb
124 ffff8801f44ec280: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
125 >ffff8801f44ec300: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 03
127 ffff8801f44ec380: fc fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb
128 ffff8801f44ec400: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
129 ==================================================================
131 The header of the report provides a short summary of what kind of bug happened
132 and what kind of access caused it. It's followed by a stack trace of the bad
133 access, a stack trace of where the accessed memory was allocated (in case bad
134 access happens on a slab object), and a stack trace of where the object was
135 freed (in case of a use-after-free bug report). Next comes a description of
136 the accessed slab object and information about the accessed memory page.
138 In the last section the report shows memory state around the accessed address.
139 Reading this part requires some understanding of how KASAN works.
141 The state of each 8 aligned bytes of memory is encoded in one shadow byte.
142 Those 8 bytes can be accessible, partially accessible, freed or be a redzone.
143 We use the following encoding for each shadow byte: 0 means that all 8 bytes
144 of the corresponding memory region are accessible; number N (1 <= N <= 7) means
145 that the first N bytes are accessible, and other (8 - N) bytes are not;
146 any negative value indicates that the entire 8-byte word is inaccessible.
147 We use different negative values to distinguish between different kinds of
148 inaccessible memory like redzones or freed memory (see mm/kasan/kasan.h).
150 In the report above the arrows point to the shadow byte 03, which means that
151 the accessed address is partially accessible.
153 For tag-based KASAN this last report section shows the memory tags around the
154 accessed address (see Implementation details section).
157 Implementation details
158 ----------------------
163 From a high level, our approach to memory error detection is similar to that
164 of kmemcheck: use shadow memory to record whether each byte of memory is safe
165 to access, and use compile-time instrumentation to insert checks of shadow
166 memory on each memory access.
168 Generic KASAN dedicates 1/8th of kernel memory to its shadow memory (e.g. 16TB
169 to cover 128TB on x86_64) and uses direct mapping with a scale and offset to
170 translate a memory address to its corresponding shadow address.
172 Here is the function which translates an address to its corresponding shadow
175 static inline void *kasan_mem_to_shadow(const void *addr)
177 return ((unsigned long)addr >> KASAN_SHADOW_SCALE_SHIFT)
178 + KASAN_SHADOW_OFFSET;
181 where ``KASAN_SHADOW_SCALE_SHIFT = 3``.
183 Compile-time instrumentation is used to insert memory access checks. Compiler
184 inserts function calls (__asan_load*(addr), __asan_store*(addr)) before each
185 memory access of size 1, 2, 4, 8 or 16. These functions check whether memory
186 access is valid or not by checking corresponding shadow memory.
188 GCC 5.0 has possibility to perform inline instrumentation. Instead of making
189 function calls GCC directly inserts the code to check the shadow memory.
190 This option significantly enlarges kernel but it gives x1.1-x2 performance
191 boost over outline instrumented kernel.
193 Generic KASAN prints up to 2 call_rcu() call stacks in reports, the last one
194 and the second to last.
196 Software tag-based KASAN
197 ~~~~~~~~~~~~~~~~~~~~~~~~
199 Tag-based KASAN uses the Top Byte Ignore (TBI) feature of modern arm64 CPUs to
200 store a pointer tag in the top byte of kernel pointers. Like generic KASAN it
201 uses shadow memory to store memory tags associated with each 16-byte memory
202 cell (therefore it dedicates 1/16th of the kernel memory for shadow memory).
204 On each memory allocation tag-based KASAN generates a random tag, tags the
205 allocated memory with this tag, and embeds this tag into the returned pointer.
206 Software tag-based KASAN uses compile-time instrumentation to insert checks
207 before each memory access. These checks make sure that tag of the memory that
208 is being accessed is equal to tag of the pointer that is used to access this
209 memory. In case of a tag mismatch tag-based KASAN prints a bug report.
211 Software tag-based KASAN also has two instrumentation modes (outline, that
212 emits callbacks to check memory accesses; and inline, that performs the shadow
213 memory checks inline). With outline instrumentation mode, a bug report is
214 simply printed from the function that performs the access check. With inline
215 instrumentation a brk instruction is emitted by the compiler, and a dedicated
216 brk handler is used to print bug reports.
218 A potential expansion of this mode is a hardware tag-based mode, which would
219 use hardware memory tagging support instead of compiler instrumentation and
220 manual shadow memory manipulation.
222 What memory accesses are sanitised by KASAN?
223 --------------------------------------------
225 The kernel maps memory in a number of different parts of the address
226 space. This poses something of a problem for KASAN, which requires
227 that all addresses accessed by instrumented code have a valid shadow
230 The range of kernel virtual addresses is large: there is not enough
231 real memory to support a real shadow region for every address that
232 could be accessed by the kernel.
237 By default, architectures only map real memory over the shadow region
238 for the linear mapping (and potentially other small areas). For all
239 other areas - such as vmalloc and vmemmap space - a single read-only
240 page is mapped over the shadow area. This read-only shadow page
241 declares all memory accesses as permitted.
243 This presents a problem for modules: they do not live in the linear
244 mapping, but in a dedicated module space. By hooking in to the module
245 allocator, KASAN can temporarily map real shadow memory to cover
246 them. This allows detection of invalid accesses to module globals, for
249 This also creates an incompatibility with ``VMAP_STACK``: if the stack
250 lives in vmalloc space, it will be shadowed by the read-only page, and
251 the kernel will fault when trying to set up the shadow data for stack
257 With ``CONFIG_KASAN_VMALLOC``, KASAN can cover vmalloc space at the
258 cost of greater memory usage. Currently this is only supported on x86.
260 This works by hooking into vmalloc and vmap, and dynamically
261 allocating real shadow memory to back the mappings.
263 Most mappings in vmalloc space are small, requiring less than a full
264 page of shadow space. Allocating a full shadow page per mapping would
265 therefore be wasteful. Furthermore, to ensure that different mappings
266 use different shadow pages, mappings would have to be aligned to
267 ``KASAN_SHADOW_SCALE_SIZE * PAGE_SIZE``.
269 Instead, we share backing space across multiple mappings. We allocate
270 a backing page when a mapping in vmalloc space uses a particular page
271 of the shadow region. This page can be shared by other vmalloc
274 We hook in to the vmap infrastructure to lazily clean up unused shadow
277 To avoid the difficulties around swapping mappings around, we expect
278 that the part of the shadow region that covers the vmalloc space will
279 not be covered by the early shadow page, but will be left
280 unmapped. This will require changes in arch-specific code.
282 This allows ``VMAP_STACK`` support on x86, and can simplify support of
283 architectures that do not have a fixed module region.