[PATCH 0/2] Fix some incompatibilites between KASAN and FORTIFY_SOURCE

5 views
Skip to first unread message

Daniel Axtens

unread,
Jan 15, 2020, 1:37:32 AM1/15/20
to linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, Daniel Axtens
3 KASAN self-tests fail on a kernel with both KASAN and FORTIFY_SOURCE:
memchr, memcmp and strlen. I have observed this on x86 and powerpc.

When FORTIFY_SOURCE is on, a number of functions are replaced with
fortified versions, which attempt to check the sizes of the
operands. However, these functions often directly invoke __builtin_foo()
once they have performed the fortify check.

This breaks things in 2 ways:

- the three function calls are technically dead code, and can be
eliminated. When __builtin_ versions are used, the compiler can detect
this.

- Using __builtins may bypass KASAN checks if the compiler decides to
inline it's own implementation as sequence of instructions, rather than
emit a function call that goes out to a KASAN-instrumented
implementation.

The patches address each reason in turn.

As a result, we're also able to remove a snippet of code copy-pasted
between every KASAN implementation that tries (largely unsuccessfully) to
disable FORTIFY_SOURCE under KASAN.

Daniel Axtens (2):
kasan: stop tests being eliminated as dead code with FORTIFY_SOURCE
string.h: fix incompatibility between FORTIFY_SOURCE and KASAN

arch/arm64/include/asm/string.h | 4 ---
arch/powerpc/include/asm/string.h | 4 ---
arch/s390/include/asm/string.h | 4 ---
arch/x86/include/asm/string_64.h | 4 ---
arch/xtensa/include/asm/string.h | 3 --
include/linux/string.h | 49 +++++++++++++++++++++++--------
lib/test_kasan.c | 30 ++++++++++++-------
7 files changed, 56 insertions(+), 42 deletions(-)

--
2.20.1

Daniel Axtens

unread,
Jan 15, 2020, 1:37:36 AM1/15/20
to linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, Daniel Axtens, Daniel Micay, Andrey Ryabinin, Alexander Potapenko, Dmitry Vyukov
3 KASAN self-tests fail on a kernel with both KASAN and FORTIFY_SOURCE:
memchr, memcmp and strlen.

When FORTIFY_SOURCE is on, a number of functions are replaced with
fortified versions, which attempt to check the sizes of the operands.
However, these functions often directly invoke __builtin_foo() once they
have performed the fortify check. The compiler can detect that the results
of these functions are not used, and knows that they have no other side
effects, and so can eliminate them as dead code.

Why are only memchr, memcmp and strlen affected?
================================================

Of string and string-like functions, kasan_test tests:

* strchr -> not affected, no fortified version
* strrchr -> likewise
* strcmp -> likewise
* strncmp -> likewise

* strnlen -> not affected, the fortify source implementation calls the
underlying strnlen implementation which is instrumented, not
a builtin

* strlen -> affected, the fortify souce implementation calls a __builtin
version which the compiler can determine is dead.

* memchr -> likewise
* memcmp -> likewise

* memset -> not affected, the compiler knows that memset writes to its
first argument and therefore is not dead.

Why does this not affect the functions normally?
================================================

In string.h, these functions are not marked as __pure, so the compiler
cannot know that they do not have side effects. If relevant functions are
marked as __pure in string.h, we see the following warnings and the
functions are elided:

lib/test_kasan.c: In function ‘kasan_memchr’:
lib/test_kasan.c:606:2: warning: statement with no effect [-Wunused-value]
memchr(ptr, '1', size + 1);
^~~~~~~~~~~~~~~~~~~~~~~~~~
lib/test_kasan.c: In function ‘kasan_memcmp’:
lib/test_kasan.c:622:2: warning: statement with no effect [-Wunused-value]
memcmp(ptr, arr, size+1);
^~~~~~~~~~~~~~~~~~~~~~~~
lib/test_kasan.c: In function ‘kasan_strings’:
lib/test_kasan.c:645:2: warning: statement with no effect [-Wunused-value]
strchr(ptr, '1');
^~~~~~~~~~~~~~~~
...

This annotation would make sense to add and could be added at any point, so
the behaviour of test_kasan.c should change.

The fix
=======

Make all the functions that are pure write their results to a global,
which makes them live. The strlen and memchr tests now pass.

The memcmp test still fails to trigger, which is addressed in the next
patch.

Cc: Daniel Micay <danie...@gmail.com>
Cc: Andrey Ryabinin <arya...@virtuozzo.com>
Cc: Alexander Potapenko <gli...@google.com>
Cc: Dmitry Vyukov <dvy...@google.com>
Fixes: 0c96350a2d2f ("lib/test_kasan.c: add tests for several string/memory API functions")
Signed-off-by: Daniel Axtens <d...@axtens.net>
---
lib/test_kasan.c | 30 +++++++++++++++++++-----------
1 file changed, 19 insertions(+), 11 deletions(-)

diff --git a/lib/test_kasan.c b/lib/test_kasan.c
index 328d33beae36..58a8cef0d7a2 100644
--- a/lib/test_kasan.c
+++ b/lib/test_kasan.c
@@ -23,6 +23,14 @@

#include <asm/page.h>

+/*
+ * We assign some test results to these globals to make sure the tests
+ * are not eliminated as dead code.
+ */
+
+int int_result;
+void *ptr_result;
+
/*
* Note: test functions are marked noinline so that their names appear in
* reports.
@@ -603,7 +611,7 @@ static noinline void __init kasan_memchr(void)
if (!ptr)
return;

- memchr(ptr, '1', size + 1);
+ ptr_result = memchr(ptr, '1', size + 1);
kfree(ptr);
}

@@ -618,8 +626,7 @@ static noinline void __init kasan_memcmp(void)
if (!ptr)
return;

- memset(arr, 0, sizeof(arr));
- memcmp(ptr, arr, size+1);
+ int_result = memcmp(ptr, arr, size + 1);
kfree(ptr);
}

@@ -642,22 +649,22 @@ static noinline void __init kasan_strings(void)
* will likely point to zeroed byte.
*/
ptr += 16;
- strchr(ptr, '1');
+ ptr_result = strchr(ptr, '1');

pr_info("use-after-free in strrchr\n");
- strrchr(ptr, '1');
+ ptr_result = strrchr(ptr, '1');

pr_info("use-after-free in strcmp\n");
- strcmp(ptr, "2");
+ int_result = strcmp(ptr, "2");

pr_info("use-after-free in strncmp\n");
- strncmp(ptr, "2", 1);
+ int_result = strncmp(ptr, "2", 1);

pr_info("use-after-free in strlen\n");
- strlen(ptr);
+ int_result = strlen(ptr);

pr_info("use-after-free in strnlen\n");
- strnlen(ptr, 1);
+ int_result = strnlen(ptr, 1);
}

static noinline void __init kasan_bitops(void)
@@ -724,11 +731,12 @@ static noinline void __init kasan_bitops(void)
__test_and_change_bit(BITS_PER_LONG + BITS_PER_BYTE, bits);

pr_info("out-of-bounds in test_bit\n");
- (void)test_bit(BITS_PER_LONG + BITS_PER_BYTE, bits);
+ int_result = test_bit(BITS_PER_LONG + BITS_PER_BYTE, bits);

#if defined(clear_bit_unlock_is_negative_byte)
pr_info("out-of-bounds in clear_bit_unlock_is_negative_byte\n");
- clear_bit_unlock_is_negative_byte(BITS_PER_LONG + BITS_PER_BYTE, bits);
+ int_result = clear_bit_unlock_is_negative_byte(BITS_PER_LONG +
+ BITS_PER_BYTE, bits);
#endif
kfree(bits);
}
--
2.20.1

Daniel Axtens

unread,
Jan 15, 2020, 1:37:40 AM1/15/20
to linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, Daniel Axtens, Daniel Micay, Andrey Ryabinin, Alexander Potapenko, Dmitry Vyukov
The memcmp KASAN self-test fails on a kernel with both KASAN and
FORTIFY_SOURCE.

When FORTIFY_SOURCE is on, a number of functions are replaced with
fortified versions, which attempt to check the sizes of the operands.
However, these functions often directly invoke __builtin_foo() once they
have performed the fortify check. Using __builtins may bypass KASAN
checks if the compiler decides to inline it's own implementation as
sequence of instructions, rather than emit a function call that goes out
to a KASAN-instrumented implementation.

Why is only memcmp affected?
============================

Of the string and string-like functions that kasan_test tests, only memcmp
is replaced by an inline sequence of instructions in my testing on x86 with
gcc version 9.2.1 20191008 (Ubuntu 9.2.1-9ubuntu2).

I believe this is due to compiler heuristics. For example, if I annotate
kmalloc calls with the alloc_size annotation (and disable some fortify
compile-time checking!), the compiler will replace every memset except the
one in kmalloc_uaf_memset with inline instructions. (I have some WIP
patches to add this annotation.)

Does this affect other functions in string.h?
=============================================

Yes. Anything that uses __builtin_* rather than __real_* could be
affected. This looks like:

- strncpy
- strcat
- strlen
- strlcpy maybe, under some circumstances?
- strncat under some circumstances
- memset
- memcpy
- memmove
- memcmp (as noted)
- memchr
- strcpy

Whether a function call is emitted always depends on the compiler. Most
bugs should get caught by FORTIFY_SOURCE, but the missed memcmp test shows
that this is not always the case.

Isn't FORTIFY_SOURCE disabled with KASAN?
========================================-

The string headers on all arches supporting KASAN disable fortify with
kasan, but only when address sanitisation is _also_ disabled. For example
from x86:

#if defined(CONFIG_KASAN) && !defined(__SANITIZE_ADDRESS__)
/*
* For files that are not instrumented (e.g. mm/slub.c) we
* should use not instrumented version of mem* functions.
*/
#define memcpy(dst, src, len) __memcpy(dst, src, len)
#define memmove(dst, src, len) __memmove(dst, src, len)
#define memset(s, c, n) __memset(s, c, n)

#ifndef __NO_FORTIFY
#define __NO_FORTIFY /* FORTIFY_SOURCE uses __builtin_memcpy, etc. */
#endif

#endif

This comes from commit 6974f0c4555e ("include/linux/string.h: add the
option of fortified string.h functions"), and doesn't work when KASAN is
enabled and the file is supposed to be sanitised - as with test_kasan.c

I'm pretty sure this is backwards: we shouldn't be using __builtin_memcpy
when we have a KASAN instrumented file, but we can use __builtin_* - and in
many cases all fortification - in files where we don't have
instrumentation.

What is correct behaviour?
==========================

Firstly, there is some overlap between fortification and KASAN: both
provide some level of _runtime_ checking. Only fortify provides
compile-time checking.

KASAN and fortify can pick up different things at runtime:

- Some fortify functions, notably the string functions, could easily be
modified to consider sub-object sizes (e.g. members within a struct),
and I have some WIP patches to do this. KASAN cannot detect these
because it cannot insert poision between members of a struct.

- KASAN can detect many over-reads/over-writes when the sizes of both
operands are unknown, which fortify cannot.

So there are a couple of options:

1) Flip the test: disable fortify in santised files and enable it in
unsanitised files. This at least stops us missing KASAN checking, but
we lose the fortify checking.

2) Make the fortify code always call out to real versions. Do this only
for KASAN, for fear of losing the inlining opportunities we get from
__builtin_*.

(We can't use kasan_check_{read,write}: because the fortify functions are
_extern inline_, you can't include _static_ inline functions without a
compiler warning. kasan_check_{read,write} are static inline so we can't
use them even when they would otherwise be suitable.)

Take approach 2 and call out to real versions when KASAN is enabled.

Use __underlying_foo to distinguish from __real_foo: __real_foo always
refers to the kernel's implementation of foo, __underlying_foo could be
either the kernel implementation or the __builtin_foo implementation.

Remove all the attempted disablement code in arch string headers.

This makes all the tests succeed with FORTIFY_SOURCE enabled.

Cc: Daniel Micay <danie...@gmail.com>
Cc: Andrey Ryabinin <arya...@virtuozzo.com>
Cc: Alexander Potapenko <gli...@google.com>
Cc: Dmitry Vyukov <dvy...@google.com>
Fixes: 6974f0c4555e ("include/linux/string.h: add the option of fortified string.h functions")
Signed-off-by: Daniel Axtens <d...@axtens.net>

---

Dmitry, this might cause a few new syzkaller splats - I first picked it up
building from a syskaller config. Or it might not, it just depends what gets
replaced with an inline sequence of instructions.

checkpatch complains about some over-long lines, happy to change the format
if anyone has better ideas for how to lay it out.
---
arch/arm64/include/asm/string.h | 4 ---
arch/powerpc/include/asm/string.h | 4 ---
arch/s390/include/asm/string.h | 4 ---
arch/x86/include/asm/string_64.h | 4 ---
arch/xtensa/include/asm/string.h | 3 --
include/linux/string.h | 49 +++++++++++++++++++++++--------
6 files changed, 37 insertions(+), 31 deletions(-)

diff --git a/arch/arm64/include/asm/string.h b/arch/arm64/include/asm/string.h
index b31e8e87a0db..eafb2c4771fc 100644
--- a/arch/arm64/include/asm/string.h
+++ b/arch/arm64/include/asm/string.h
@@ -59,10 +59,6 @@ void memcpy_flushcache(void *dst, const void *src, size_t cnt);
#define memmove(dst, src, len) __memmove(dst, src, len)
#define memset(s, c, n) __memset(s, c, n)

-#ifndef __NO_FORTIFY
-#define __NO_FORTIFY /* FORTIFY_SOURCE uses __builtin_memcpy, etc. */
-#endif
-
#endif

#endif
diff --git a/arch/powerpc/include/asm/string.h b/arch/powerpc/include/asm/string.h
index b72692702f35..952c5934596b 100644
--- a/arch/powerpc/include/asm/string.h
+++ b/arch/powerpc/include/asm/string.h
@@ -43,10 +43,6 @@ void *__memmove(void *to, const void *from, __kernel_size_t n);
#define memmove(dst, src, len) __memmove(dst, src, len)
#define memset(s, c, n) __memset(s, c, n)

-#ifndef __NO_FORTIFY
-#define __NO_FORTIFY /* FORTIFY_SOURCE uses __builtin_memcpy, etc. */
-#endif
-
#endif

#ifdef CONFIG_PPC64
diff --git a/arch/s390/include/asm/string.h b/arch/s390/include/asm/string.h
index 4c0690fc5167..e0b66d8c89a1 100644
--- a/arch/s390/include/asm/string.h
+++ b/arch/s390/include/asm/string.h
@@ -75,10 +75,6 @@ extern void *__memmove(void *dest, const void *src, size_t n);

#define __no_sanitize_prefix_strfunc(x) __##x

-#ifndef __NO_FORTIFY
-#define __NO_FORTIFY /* FORTIFY_SOURCE uses __builtin_memcpy, etc. */
-#endif
-
#else
#define __no_sanitize_prefix_strfunc(x) x
#endif /* defined(CONFIG_KASAN) && !defined(__SANITIZE_ADDRESS__) */
diff --git a/arch/x86/include/asm/string_64.h b/arch/x86/include/asm/string_64.h
index 75314c3dbe47..ec63d11e1f04 100644
--- a/arch/x86/include/asm/string_64.h
+++ b/arch/x86/include/asm/string_64.h
@@ -76,10 +76,6 @@ int strcmp(const char *cs, const char *ct);
#define memmove(dst, src, len) __memmove(dst, src, len)
#define memset(s, c, n) __memset(s, c, n)

-#ifndef __NO_FORTIFY
-#define __NO_FORTIFY /* FORTIFY_SOURCE uses __builtin_memcpy, etc. */
-#endif
-
#endif

#define __HAVE_ARCH_MEMCPY_MCSAFE 1
diff --git a/arch/xtensa/include/asm/string.h b/arch/xtensa/include/asm/string.h
index 89b51a0c752f..8cf04c5a33fb 100644
--- a/arch/xtensa/include/asm/string.h
+++ b/arch/xtensa/include/asm/string.h
@@ -132,9 +132,6 @@ extern void *__memmove(void *__dest, __const__ void *__src, size_t __n);
#define memmove(dst, src, len) __memmove(dst, src, len)
#define memset(s, c, n) __memset(s, c, n)

-#ifndef __NO_FORTIFY
-#define __NO_FORTIFY /* FORTIFY_SOURCE uses __builtin_memcpy, etc. */
-#endif
#endif

#endif /* _XTENSA_STRING_H */
diff --git a/include/linux/string.h b/include/linux/string.h
index 3b8e8b12dd37..4364c106355e 100644
--- a/include/linux/string.h
+++ b/include/linux/string.h
@@ -317,6 +317,31 @@ void __read_overflow3(void) __compiletime_error("detected read beyond size of ob
void __write_overflow(void) __compiletime_error("detected write beyond size of object passed as 1st parameter");

#if !defined(__NO_FORTIFY) && defined(__OPTIMIZE__) && defined(CONFIG_FORTIFY_SOURCE)
+
+#ifdef CONFIG_KASAN
+extern void *__underlying_memchr(const void *p, int c, __kernel_size_t size) __RENAME(memchr);
+extern int __underlying_memcmp(const void *p, const void *q, __kernel_size_t size) __RENAME(memcmp);
+extern void *__underlying_memcpy(void *p, const void *q, __kernel_size_t size) __RENAME(memcpy);
+extern void *__underlying_memmove(void *p, const void *q, __kernel_size_t size) __RENAME(memmove);
+extern void *__underlying_memset(void *p, int c, __kernel_size_t size) __RENAME(memset);
+extern char *__underlying_strcat(char *p, const char *q) __RENAME(strcat);
+extern char *__underlying_strcpy(char *p, const char *q) __RENAME(strcpy);
+extern __kernel_size_t __underlying_strlen(const char *p) __RENAME(strlen);
+extern char *__underlying_strncat(char *p, const char *q, __kernel_size_t count) __RENAME(strncat);
+extern char *__underlying_strncpy(char *p, const char *q, __kernel_size_t size) __RENAME(strncpy);
+#else
+#define __underlying_memchr __builtin_memchr
+#define __underlying_memcmp __builtin_memcmp
+#define __underlying_memcpy __builtin_memcpy
+#define __underlying_memmove __builtin_memmove
+#define __underlying_memset __builtin_memset
+#define __underlying_strcat __builtin_strcat
+#define __underlying_strcpy __builtin_strcpy
+#define __underlying_strlen __builtin_strlen
+#define __underlying_strncat __builtin_strncat
+#define __underlying_strncpy __builtin_strncpy
+#endif
+
__FORTIFY_INLINE char *strncpy(char *p, const char *q, __kernel_size_t size)
{
size_t p_size = __builtin_object_size(p, 0);
@@ -324,14 +349,14 @@ __FORTIFY_INLINE char *strncpy(char *p, const char *q, __kernel_size_t size)
__write_overflow();
if (p_size < size)
fortify_panic(__func__);
- return __builtin_strncpy(p, q, size);
+ return __underlying_strncpy(p, q, size);
}

__FORTIFY_INLINE char *strcat(char *p, const char *q)
{
size_t p_size = __builtin_object_size(p, 0);
if (p_size == (size_t)-1)
- return __builtin_strcat(p, q);
+ return __underlying_strcat(p, q);
if (strlcat(p, q, p_size) >= p_size)
fortify_panic(__func__);
return p;
@@ -345,7 +370,7 @@ __FORTIFY_INLINE __kernel_size_t strlen(const char *p)
/* Work around gcc excess stack consumption issue */
if (p_size == (size_t)-1 ||
(__builtin_constant_p(p[p_size - 1]) && p[p_size - 1] == '\0'))
- return __builtin_strlen(p);
+ return __underlying_strlen(p);
ret = strnlen(p, p_size);
if (p_size <= ret)
fortify_panic(__func__);
@@ -378,7 +403,7 @@ __FORTIFY_INLINE size_t strlcpy(char *p, const char *q, size_t size)
__write_overflow();
if (len >= p_size)
fortify_panic(__func__);
- __builtin_memcpy(p, q, len);
+ __underlying_memcpy(p, q, len);
p[len] = '\0';
}
return ret;
@@ -391,12 +416,12 @@ __FORTIFY_INLINE char *strncat(char *p, const char *q, __kernel_size_t count)
size_t p_size = __builtin_object_size(p, 0);
size_t q_size = __builtin_object_size(q, 0);
if (p_size == (size_t)-1 && q_size == (size_t)-1)
- return __builtin_strncat(p, q, count);
+ return __underlying_strncat(p, q, count);
p_len = strlen(p);
copy_len = strnlen(q, count);
if (p_size < p_len + copy_len + 1)
fortify_panic(__func__);
- __builtin_memcpy(p + p_len, q, copy_len);
+ __underlying_memcpy(p + p_len, q, copy_len);
p[p_len + copy_len] = '\0';
return p;
}
@@ -408,7 +433,7 @@ __FORTIFY_INLINE void *memset(void *p, int c, __kernel_size_t size)
__write_overflow();
if (p_size < size)
fortify_panic(__func__);
- return __builtin_memset(p, c, size);
+ return __underlying_memset(p, c, size);
}

__FORTIFY_INLINE void *memcpy(void *p, const void *q, __kernel_size_t size)
@@ -423,7 +448,7 @@ __FORTIFY_INLINE void *memcpy(void *p, const void *q, __kernel_size_t size)
}
if (p_size < size || q_size < size)
fortify_panic(__func__);
- return __builtin_memcpy(p, q, size);
+ return __underlying_memcpy(p, q, size);
}

__FORTIFY_INLINE void *memmove(void *p, const void *q, __kernel_size_t size)
@@ -438,7 +463,7 @@ __FORTIFY_INLINE void *memmove(void *p, const void *q, __kernel_size_t size)
}
if (p_size < size || q_size < size)
fortify_panic(__func__);
- return __builtin_memmove(p, q, size);
+ return __underlying_memmove(p, q, size);
}

extern void *__real_memscan(void *, int, __kernel_size_t) __RENAME(memscan);
@@ -464,7 +489,7 @@ __FORTIFY_INLINE int memcmp(const void *p, const void *q, __kernel_size_t size)
}
if (p_size < size || q_size < size)
fortify_panic(__func__);
- return __builtin_memcmp(p, q, size);
+ return __underlying_memcmp(p, q, size);
}

__FORTIFY_INLINE void *memchr(const void *p, int c, __kernel_size_t size)
@@ -474,7 +499,7 @@ __FORTIFY_INLINE void *memchr(const void *p, int c, __kernel_size_t size)
__read_overflow();
if (p_size < size)
fortify_panic(__func__);
- return __builtin_memchr(p, c, size);
+ return __underlying_memchr(p, c, size);
}

void *__real_memchr_inv(const void *s, int c, size_t n) __RENAME(memchr_inv);
@@ -505,7 +530,7 @@ __FORTIFY_INLINE char *strcpy(char *p, const char *q)
size_t p_size = __builtin_object_size(p, 0);
size_t q_size = __builtin_object_size(q, 0);
if (p_size == (size_t)-1 && q_size == (size_t)-1)
- return __builtin_strcpy(p, q);
+ return __underlying_strcpy(p, q);
memcpy(p, q, strlen(q) + 1);
return p;
}
--
2.20.1

Dmitry Vyukov

unread,
Jan 15, 2020, 9:43:26 AM1/15/20
to Daniel Axtens, LKML, Linux-MM, kasan-dev, linuxppc-dev, Linux ARM, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, Daniel Micay, Andrey Ryabinin, Alexander Potapenko
These are globals, but are not static and don't have kasan_ prefix.
But I guess this does not matter for modules?
Otherwise:

Reviewed-by: Dmitry Vyukov <dvy...@google.com>

Christophe Leroy

unread,
Jan 15, 2020, 9:47:12 AM1/15/20
to Dmitry Vyukov, Daniel Axtens, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, LKML, kasan-dev, Linux-MM, Daniel Micay, Alexander Potapenko, Andrey Ryabinin, linuxppc-dev, Linux ARM
I think if you make them static, GCC will see they aren't used and will
eliminate everything still ?

Christophe

Dmitry Vyukov

unread,
Jan 15, 2020, 9:57:04 AM1/15/20
to Daniel Axtens, LKML, Linux-MM, kasan-dev, linuxppc-dev, Linux ARM, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, Daniel Micay, Andrey Ryabinin, Alexander Potapenko
On Wed, Jan 15, 2020 at 7:37 AM Daniel Axtens <d...@axtens.net> wrote:
>
Hi Daniel,

Thanks for addressing this. And special kudos for description detail level! :)

Phew, this layering of checking tools is a bit messy...

> I'm pretty sure this is backwards: we shouldn't be using __builtin_memcpy
> when we have a KASAN instrumented file, but we can use __builtin_* - and in
> many cases all fortification - in files where we don't have
> instrumentation.

I think if we use __builtin_* in a non-instrumented file, the compiler
can emit a call to normal mem* function which will be intercepted by
kasan and we will get instrumentation in a file which should not be
instrumented. Moreover this behavior will depend on optimization level
and compiler internals.
But as far as I see this does not affect any of the following and the
code change.



> What is correct behaviour?
> ==========================
>
> Firstly, there is some overlap between fortification and KASAN: both
> provide some level of _runtime_ checking. Only fortify provides
> compile-time checking.
>
> KASAN and fortify can pick up different things at runtime:
>
> - Some fortify functions, notably the string functions, could easily be
> modified to consider sub-object sizes (e.g. members within a struct),
> and I have some WIP patches to do this. KASAN cannot detect these
> because it cannot insert poision between members of a struct.
>
> - KASAN can detect many over-reads/over-writes when the sizes of both
> operands are unknown, which fortify cannot.
>
> So there are a couple of options:
>
> 1) Flip the test: disable fortify in santised files and enable it in
> unsanitised files. This at least stops us missing KASAN checking, but
> we lose the fortify checking.
>
> 2) Make the fortify code always call out to real versions. Do this only
> for KASAN, for fear of losing the inlining opportunities we get from
> __builtin_*.
>
> (We can't use kasan_check_{read,write}: because the fortify functions are
> _extern inline_, you can't include _static_ inline functions without a
> compiler warning. kasan_check_{read,write} are static inline so we can't
> use them even when they would otherwise be suitable.)
>
> Take approach 2 and call out to real versions when KASAN is enabled.

I support option 2.
For KASAN build we don't care about inlining/performance that much,
getting it to work reliably and with reasonable complexity is more
important.
And it's better to leave prod build as it is now (proving that any
change is harmless is impossible).
arch headers do:

#if defined(CONFIG_KASAN) && !defined(__SANITIZE_ADDRESS__)
#define memcpy(dst, src, len) __memcpy(dst, src, len)
...

to disable instrumentation. Does this still work with this change?
Previously they disabled fortify. What happens now? Will define of
memcpy to __memcpy also affect __RENAME(memcpy), so that
__underlying_memcpy will be an alias to __memcpy?



> +extern int __underlying_memcmp(const void *p, const void *q, __kernel_size_t size) __RENAME(memcmp);

All of these macros are leaking from the header file. Tomorrow we will
discover __underlying_memcpy uses somewhere in the wild, which will
not making understanding what actually happens simpler :)
Perhaps undef all of them at the bottom?

Dmitry Vyukov

unread,
Jan 15, 2020, 9:57:49 AM1/15/20
to Christophe Leroy, Daniel Axtens, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, LKML, kasan-dev, Linux-MM, Daniel Micay, Alexander Potapenko, Andrey Ryabinin, linuxppc-dev, Linux ARM
static volatile? :)

Daniel Axtens

unread,
Jan 15, 2020, 11:59:28 PM1/15/20
to Dmitry Vyukov, LKML, Linux-MM, kasan-dev, linuxppc-dev, Linux ARM, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, Daniel Micay, Andrey Ryabinin, Alexander Potapenko
mmm OK - you are right, when I consider this and your other point...

>> #if !defined(__NO_FORTIFY) && defined(__OPTIMIZE__) && defined(CONFIG_FORTIFY_SOURCE)
>> +
>> +#ifdef CONFIG_KASAN
>> +extern void *__underlying_memchr(const void *p, int c, __kernel_size_t size) __RENAME(memchr);
>
>
> arch headers do:
>
> #if defined(CONFIG_KASAN) && !defined(__SANITIZE_ADDRESS__)
> #define memcpy(dst, src, len) __memcpy(dst, src, len)
> ...
>
> to disable instrumentation. Does this still work with this change?
> Previously they disabled fortify. What happens now? Will define of
> memcpy to __memcpy also affect __RENAME(memcpy), so that
> __underlying_memcpy will be an alias to __memcpy?

This is a good question. It's a really intricate set of interactions!!

Between these two things, I think I'm going to just drop the removal of
architecture changes, which means that fortify will continue to be
disabled for files that disable KASAN sanitisation. It's just too
complicated to reason through and satisfy myself that we're not going to
get weird bugs, and the payoff is really small.

>> +extern int __underlying_memcmp(const void *p, const void *q, __kernel_size_t size) __RENAME(memcmp);
>
> All of these macros are leaking from the header file. Tomorrow we will
> discover __underlying_memcpy uses somewhere in the wild, which will
> not making understanding what actually happens simpler :)
> Perhaps undef all of them at the bottom?

I can't stop the function definitions from leaking, but I can stop the
defines from leaking, which means we will catch any uses outside this
block in a FORITY_SOURCE && !KASAN build. I've fixed this for v2.

Regards,
Daniel

Daniel Axtens

unread,
Jan 16, 2020, 12:34:27 AM1/16/20
to Dmitry Vyukov, Christophe Leroy, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, LKML, kasan-dev, Linux-MM, Daniel Micay, Alexander Potapenko, Andrey Ryabinin, linuxppc-dev, Linux ARM
>> >> +/*
>> >> + * We assign some test results to these globals to make sure the tests
>> >> + * are not eliminated as dead code.
>> >> + */
>> >> +
>> >> +int int_result;
>> >> +void *ptr_result;
>> >
>> > These are globals, but are not static and don't have kasan_ prefix.
>> > But I guess this does not matter for modules?
>> > Otherwise:
>> >
>> > Reviewed-by: Dmitry Vyukov <dvy...@google.com>
>> >
>>
>> I think if you make them static, GCC will see they aren't used and will
>> eliminate everything still ?
>
> static volatile? :)

Yeah so these are module globals. They'd be accessible from any other
files you linked into the module (currently there are no such
files). They're not visible outside the module because they're not
EXPORTed.

Making them static does lead to them getting eliminated, and 'static
volatile' seems both gross and like something checkpatch would complain
about. I'll leave them as they are but stick a kasan_ prefix on them
just for the additional tidiness.

Regards,
Daniel

Dmitry Vyukov

unread,
Jan 16, 2020, 12:47:25 AM1/16/20
to Daniel Axtens, LKML, Linux-MM, kasan-dev, linuxppc-dev, Linux ARM, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, Daniel Micay, Andrey Ryabinin, Alexander Potapenko
Sounds good to me. We don't need to solve all of the world 's problems
at once :)

> >> +extern int __underlying_memcmp(const void *p, const void *q, __kernel_size_t size) __RENAME(memcmp);
> >
> > All of these macros are leaking from the header file. Tomorrow we will
> > discover __underlying_memcpy uses somewhere in the wild, which will
> > not making understanding what actually happens simpler :)
> > Perhaps undef all of them at the bottom?
>
> I can't stop the function definitions from leaking, but I can stop the
> defines from leaking, which means we will catch any uses outside this
> block in a FORITY_SOURCE && !KASAN build. I've fixed this for v2.

I think it's good enough and a good practice to undef local macros.
> --
> You received this message because you are subscribed to the Google Groups "kasan-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to kasan-dev+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/kasan-dev/8736cgkndh.fsf%40dja-thinkpad.axtens.net.

Daniel Axtens

unread,
Jan 16, 2020, 1:26:32 AM1/16/20
to linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, dvy...@google.com, christop...@c-s.fr, Daniel Axtens
3 KASAN self-tests fail on a kernel with both KASAN and FORTIFY_SOURCE:
memchr, memcmp and strlen. I have observed this on x86 and powerpc.

When FORTIFY_SOURCE is on, a number of functions are replaced with
fortified versions, which attempt to check the sizes of the
operands. However, these functions often directly invoke __builtin_foo()
once they have performed the fortify check.

This breaks things in 2 ways:

- the three function calls are technically dead code, and can be
eliminated. When __builtin_ versions are used, the compiler can detect
this.

- Using __builtins may bypass KASAN checks if the compiler decides to
inline it's own implementation as sequence of instructions, rather than
emit a function call that goes out to a KASAN-instrumented
implementation.

The patches address each reason in turn. Finally, test_memcmp used a
stack array without explicit initialisation, which can sometimes break
too, so fix that up.

v2: - some cleanups, don't mess with arch code as I missed some wrinkles.
- add stack array init (patch 3)

Daniel Axtens (3):
kasan: stop tests being eliminated as dead code with FORTIFY_SOURCE
string.h: fix incompatibility between FORTIFY_SOURCE and KASAN
kasan: initialise array in kasan_memcmp test

include/linux/string.h | 60 +++++++++++++++++++++++++++++++++---------
lib/test_kasan.c | 32 +++++++++++++---------
2 files changed, 68 insertions(+), 24 deletions(-)

--
2.20.1

Daniel Axtens

unread,
Jan 16, 2020, 1:26:36 AM1/16/20
to linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, dvy...@google.com, christop...@c-s.fr, Daniel Axtens, Daniel Micay, Andrey Ryabinin, Alexander Potapenko
3 KASAN self-tests fail on a kernel with both KASAN and FORTIFY_SOURCE:
memchr, memcmp and strlen.

When FORTIFY_SOURCE is on, a number of functions are replaced with
fortified versions, which attempt to check the sizes of the operands.
However, these functions often directly invoke __builtin_foo() once they
Reviewed-by: Dmitry Vyukov <dvy...@google.com>
Signed-off-by: Daniel Axtens <d...@axtens.net>

---

v2: rename variables to have kasan_ prefixes
---
lib/test_kasan.c | 30 +++++++++++++++++++-----------
1 file changed, 19 insertions(+), 11 deletions(-)

diff --git a/lib/test_kasan.c b/lib/test_kasan.c
index 328d33beae36..a130d75b9385 100644
--- a/lib/test_kasan.c
+++ b/lib/test_kasan.c
@@ -23,6 +23,14 @@

#include <asm/page.h>

+/*
+ * We assign some test results to these globals to make sure the tests
+ * are not eliminated as dead code.
+ */
+
+int kasan_int_result;
+void *kasan_ptr_result;
+
/*
* Note: test functions are marked noinline so that their names appear in
* reports.
@@ -603,7 +611,7 @@ static noinline void __init kasan_memchr(void)
if (!ptr)
return;

- memchr(ptr, '1', size + 1);
+ kasan_ptr_result = memchr(ptr, '1', size + 1);
kfree(ptr);
}

@@ -618,8 +626,7 @@ static noinline void __init kasan_memcmp(void)
if (!ptr)
return;

- memset(arr, 0, sizeof(arr));
- memcmp(ptr, arr, size+1);
+ kasan_int_result = memcmp(ptr, arr, size + 1);
kfree(ptr);
}

@@ -642,22 +649,22 @@ static noinline void __init kasan_strings(void)
* will likely point to zeroed byte.
*/
ptr += 16;
- strchr(ptr, '1');
+ kasan_ptr_result = strchr(ptr, '1');

pr_info("use-after-free in strrchr\n");
- strrchr(ptr, '1');
+ kasan_ptr_result = strrchr(ptr, '1');

pr_info("use-after-free in strcmp\n");
- strcmp(ptr, "2");
+ kasan_int_result = strcmp(ptr, "2");

pr_info("use-after-free in strncmp\n");
- strncmp(ptr, "2", 1);
+ kasan_int_result = strncmp(ptr, "2", 1);

pr_info("use-after-free in strlen\n");
- strlen(ptr);
+ kasan_int_result = strlen(ptr);

pr_info("use-after-free in strnlen\n");
- strnlen(ptr, 1);
+ kasan_int_result = strnlen(ptr, 1);
}

static noinline void __init kasan_bitops(void)
@@ -724,11 +731,12 @@ static noinline void __init kasan_bitops(void)
__test_and_change_bit(BITS_PER_LONG + BITS_PER_BYTE, bits);

pr_info("out-of-bounds in test_bit\n");
- (void)test_bit(BITS_PER_LONG + BITS_PER_BYTE, bits);
+ kasan_int_result = test_bit(BITS_PER_LONG + BITS_PER_BYTE, bits);

#if defined(clear_bit_unlock_is_negative_byte)
pr_info("out-of-bounds in clear_bit_unlock_is_negative_byte\n");
- clear_bit_unlock_is_negative_byte(BITS_PER_LONG + BITS_PER_BYTE, bits);
+ kasan_int_result = clear_bit_unlock_is_negative_byte(BITS_PER_LONG +

Daniel Axtens

unread,
Jan 16, 2020, 1:26:40 AM1/16/20
to linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, dvy...@google.com, christop...@c-s.fr, Daniel Axtens, Daniel Micay, Andrey Ryabinin, Alexander Potapenko
The memcmp KASAN self-test fails on a kernel with both KASAN and
FORTIFY_SOURCE.

When FORTIFY_SOURCE is on, a number of functions are replaced with
fortified versions, which attempt to check the sizes of the operands.
However, these functions often directly invoke __builtin_foo() once they
have performed the fortify check. Using __builtins may bypass KASAN
checks if the compiler decides to inline it's own implementation as
sequence of instructions, rather than emit a function call that goes out
to a KASAN-instrumented implementation.

I'm pretty sure this is not wrong, but not as expansive it should be:

* we shouldn't use __builtin_memcpy etc in files where we don't have
instrumentation - it could devolve into a function call to memcpy,
which will be instrumented. Rather, we should use __memcpy which
by convention is not instrumented.

* we also shouldn't be using __builtin_memcpy when we have a KASAN
instrumented file, because it could be replaced with inline asm
that will not be instrumented.

What is correct behaviour?
==========================

Firstly, there is some overlap between fortification and KASAN: both
provide some level of _runtime_ checking. Only fortify provides
compile-time checking.

KASAN and fortify can pick up different things at runtime:

- Some fortify functions, notably the string functions, could easily be
modified to consider sub-object sizes (e.g. members within a struct),
and I have some WIP patches to do this. KASAN cannot detect these
because it cannot insert poision between members of a struct.

- KASAN can detect many over-reads/over-writes when the sizes of both
operands are unknown, which fortify cannot.

So there are a couple of options:

1) Flip the test: disable fortify in santised files and enable it in
unsanitised files. This at least stops us missing KASAN checking, but
we lose the fortify checking.

2) Make the fortify code always call out to real versions. Do this only
for KASAN, for fear of losing the inlining opportunities we get from
__builtin_*.

(We can't use kasan_check_{read,write}: because the fortify functions are
_extern inline_, you can't include _static_ inline functions without a
compiler warning. kasan_check_{read,write} are static inline so we can't
use them even when they would otherwise be suitable.)

Take approach 2 and call out to real versions when KASAN is enabled.

Use __underlying_foo to distinguish from __real_foo: __real_foo always
refers to the kernel's implementation of foo, __underlying_foo could be
either the kernel implementation or the __builtin_foo implementation.

This is sometimes enough to make the memcmp test succeed with
FORTIFY_SOURCE enabled. It is at least enough to get the function call
into the module. One more fix is needed to make it reliable: see the next
patch.

Cc: Daniel Micay <danie...@gmail.com>
Cc: Andrey Ryabinin <arya...@virtuozzo.com>
Cc: Alexander Potapenko <gli...@google.com>
Cc: Dmitry Vyukov <dvy...@google.com>
Fixes: 6974f0c4555e ("include/linux/string.h: add the option of fortified string.h functions")
Signed-off-by: Daniel Axtens <d...@axtens.net>

---

v2: add #undefs, do not drop arch code: Dmitry pointed out that we _do_ want
to disable fortify in non-sanitised files because of how __builtin_memcpy
might end up as a call to regular memcpy rather than __memcpy.

Dmitry, this might cause a few new syzkaller splats - I first picked it up
building from a syskaller config. Or it might not, it just depends what gets
replaced with an inline sequence of instructions.

checkpatch complains about some over-long lines, happy to change the format
if anyone has better ideas for how to lay it out.
---
include/linux/string.h | 60 +++++++++++++++++++++++++++++++++---------
1 file changed, 48 insertions(+), 12 deletions(-)

diff --git a/include/linux/string.h b/include/linux/string.h
index 3b8e8b12dd37..18d3f7a4b2b9 100644
--- a/include/linux/string.h
+++ b/include/linux/string.h
@@ -317,6 +317,31 @@ void __read_overflow3(void) __compiletime_error("detected read beyond size of ob
void __write_overflow(void) __compiletime_error("detected write beyond size of object passed as 1st parameter");

#if !defined(__NO_FORTIFY) && defined(__OPTIMIZE__) && defined(CONFIG_FORTIFY_SOURCE)
+
+#ifdef CONFIG_KASAN
+extern void *__underlying_memchr(const void *p, int c, __kernel_size_t size) __RENAME(memchr);
+extern int __underlying_memcmp(const void *p, const void *q, __kernel_size_t size) __RENAME(memcmp);
@@ -505,11 +530,22 @@ __FORTIFY_INLINE char *strcpy(char *p, const char *q)
size_t p_size = __builtin_object_size(p, 0);
size_t q_size = __builtin_object_size(q, 0);
if (p_size == (size_t)-1 && q_size == (size_t)-1)
- return __builtin_strcpy(p, q);
+ return __underlying_strcpy(p, q);
memcpy(p, q, strlen(q) + 1);
return p;
}

+/* Don't use these outside the FORITFY_SOURCE implementation */
+#undef __underlying_memchr
+#undef __underlying_memcmp
+#undef __underlying_memcpy
+#undef __underlying_memmove
+#undef __underlying_memset
+#undef __underlying_strcat
+#undef __underlying_strcpy
+#undef __underlying_strlen
+#undef __underlying_strncat
+#undef __underlying_strncpy
#endif

/**
--
2.20.1

Daniel Axtens

unread,
Jan 16, 2020, 1:26:44 AM1/16/20
to linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, dvy...@google.com, christop...@c-s.fr, Daniel Axtens, Andrey Ryabinin, Alexander Potapenko
memcmp may bail out before accessing all the memory if the buffers
contain differing bytes. kasan_memcmp calls memcmp with a stack array.
Stack variables are not necessarily initialised (in the absence of a
compiler plugin, at least). Sometimes this causes the memcpy to bail
early thus fail to trigger kasan.

Make sure the array initialised to zero in the code.

No other test is dependent on the contents of an array on the stack.

Cc: Andrey Ryabinin <arya...@virtuozzo.com>
Cc: Alexander Potapenko <gli...@google.com>
Cc: Dmitry Vyukov <dvy...@google.com>
Signed-off-by: Daniel Axtens <d...@axtens.net>
---
lib/test_kasan.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/test_kasan.c b/lib/test_kasan.c
index a130d75b9385..519b0f259e97 100644
--- a/lib/test_kasan.c
+++ b/lib/test_kasan.c
@@ -619,7 +619,7 @@ static noinline void __init kasan_memcmp(void)
{
char *ptr;
size_t size = 24;
- int arr[9];
+ int arr[9] = {};

pr_info("out-of-bounds in memcmp\n");
ptr = kmalloc(size, GFP_KERNEL | __GFP_ZERO);
--
2.20.1

Dmitry Vyukov

unread,
Jan 16, 2020, 3:07:44 AM1/16/20
to Daniel Axtens, LKML, Linux-MM, kasan-dev, linuxppc-dev, Linux ARM, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, Christophe Leroy, Andrey Ryabinin, Alexander Potapenko
On Thu, Jan 16, 2020 at 7:26 AM Daniel Axtens <d...@axtens.net> wrote:
>
> memcmp may bail out before accessing all the memory if the buffers
> contain differing bytes. kasan_memcmp calls memcmp with a stack array.
> Stack variables are not necessarily initialised (in the absence of a
> compiler plugin, at least). Sometimes this causes the memcpy to bail
> early thus fail to trigger kasan.
>
> Make sure the array initialised to zero in the code.
>
> No other test is dependent on the contents of an array on the stack.
>
> Cc: Andrey Ryabinin <arya...@virtuozzo.com>
> Cc: Alexander Potapenko <gli...@google.com>
> Cc: Dmitry Vyukov <dvy...@google.com>
> Signed-off-by: Daniel Axtens <d...@axtens.net>

Reviewed-by: Dmitry Vyukov <dvy...@google.com>

> ---
> lib/test_kasan.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib/test_kasan.c b/lib/test_kasan.c
> index a130d75b9385..519b0f259e97 100644
> --- a/lib/test_kasan.c
> +++ b/lib/test_kasan.c
> @@ -619,7 +619,7 @@ static noinline void __init kasan_memcmp(void)
> {
> char *ptr;
> size_t size = 24;
> - int arr[9];
> + int arr[9] = {};
>
> pr_info("out-of-bounds in memcmp\n");
> ptr = kmalloc(size, GFP_KERNEL | __GFP_ZERO);
> --
> 2.20.1
>
> --
> You received this message because you are subscribed to the Google Groups "kasan-dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to kasan-dev+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/kasan-dev/20200116062625.32692-4-dja%40axtens.net.

Dmitry Vyukov

unread,
Jan 16, 2020, 3:11:42 AM1/16/20
to Daniel Axtens, LKML, Linux-MM, kasan-dev, linuxppc-dev, Linux ARM, linux-s390, linux-...@linux-xtensa.org, the arch/x86 maintainers, Christophe Leroy, Daniel Micay, Andrey Ryabinin, Alexander Potapenko
On Thu, Jan 16, 2020 at 7:26 AM Daniel Axtens <d...@axtens.net> wrote:
>
Reviewed-by: Dmitry Vyukov <dvy...@google.com>

> ---
>
> v2: add #undefs, do not drop arch code: Dmitry pointed out that we _do_ want
> to disable fortify in non-sanitised files because of how __builtin_memcpy
> might end up as a call to regular memcpy rather than __memcpy.
>
> Dmitry, this might cause a few new syzkaller splats - I first picked it up
> building from a syskaller config. Or it might not, it just depends what gets
> replaced with an inline sequence of instructions.

If you mean new true bugs, I don't think it's changing anything on top
of the hundreds of known existing open bugs:
https://syzkaller.appspot.com/upstream#open
If anything, I would say it's good to surface more true bugs.

kbuild test robot

unread,
Jan 16, 2020, 6:53:01 PM1/16/20
to Daniel Axtens, kbuil...@lists.01.org, linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, Daniel Axtens, Daniel Micay, Andrey Ryabinin, Alexander Potapenko, Dmitry Vyukov
Hi Daniel,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on linux/master]
[also build test WARNING on arm64/for-next/core powerpc/next s390/features tip/x86/core linus/master v5.5-rc6 next-20200110]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system. BTW, we also suggest to use '--base' option to specify the
base tree in git format-patch, please see https://stackoverflow.com/a/37406982]

url: https://github.com/0day-ci/linux/commits/Daniel-Axtens/Fix-some-incompatibilites-between-KASAN-and-FORTIFY_SOURCE/20200116-172838
base: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git 1522d9da40bdfe502c91163e6d769332897201fa
config: x86_64-randconfig-s1-20200116 (attached as .config)
compiler: gcc-7 (Debian 7.5.0-3) 7.5.0
reproduce:
# save the attached .config to linux build tree
make ARCH=x86_64

If you fix the issue, kindly add following tag
Reported-by: kbuild test robot <l...@intel.com>

All warnings (new ones prefixed by >>):

>> mm/kasan/common.o: warning: objtool: __kasan_slab_free()+0xea: unreachable instruction

---
0-DAY kernel test infrastructure Open Source Technology Center
https://lists.01.org/hyperkitty/list/kbuil...@lists.01.org Intel Corporation
.config.gz

Daniel Axtens

unread,
Feb 26, 2020, 6:48:32 PM2/26/20
to linux-...@vger.kernel.org, linu...@kvack.org, kasa...@googlegroups.com, linuxp...@lists.ozlabs.org, linux-ar...@lists.infradead.org, linux...@vger.kernel.org, linux-...@linux-xtensa.org, x...@kernel.org, dvy...@google.com, christop...@c-s.fr
Daniel Axtens <d...@axtens.net> writes:

> 3 KASAN self-tests fail on a kernel with both KASAN and FORTIFY_SOURCE:
> memchr, memcmp and strlen. I have observed this on x86 and powerpc.
>
> When FORTIFY_SOURCE is on, a number of functions are replaced with
> fortified versions, which attempt to check the sizes of the
> operands. However, these functions often directly invoke __builtin_foo()
> once they have performed the fortify check.
>
> This breaks things in 2 ways:
>
> - the three function calls are technically dead code, and can be
> eliminated. When __builtin_ versions are used, the compiler can detect
> this.
>
> - Using __builtins may bypass KASAN checks if the compiler decides to
> inline it's own implementation as sequence of instructions, rather than
> emit a function call that goes out to a KASAN-instrumented
> implementation.
>
> The patches address each reason in turn. Finally, test_memcmp used a
> stack array without explicit initialisation, which can sometimes break
> too, so fix that up.

Hi all,

It doesn't look like this has been picked up yet. Is there anything I
can do to help things along?

Regards,
Daniel
Reply all
Reply to author
Forward
0 new messages