[PATCH 1/4] x86: kmsan: Don't rename memintrinsics in uninstrumented files

[PATCH 1/4] x86: kmsan: Don't rename memintrinsics in uninstrumented files

[Alexander Potapenko]

KMSAN should be overriding calls to memset/memcpy/memmove and their
__builtin_ versions in instrumented files, so there is no need to
override them. In non-instrumented versions we are now required to
leave memset() and friends intact, so we cannot replace them with
__msan_XXX() functions.

Cc: Kees Cook <keescook@chromium.org>
Suggested-by: Marco Elver <elver@google.com>
Signed-off-by: Alexander Potapenko <glider@google.com>
---
 arch/x86/include/asm/string_64.h | 17 -----------------
 1 file changed, 17 deletions(-)

diff --git a/arch/x86/include/asm/string_64.h b/arch/x86/include/asm/string_64.h
index 888731ccf1f67..9be401d971a99 100644
--- a/arch/x86/include/asm/string_64.h
+++ b/arch/x86/include/asm/string_64.h
@@ -15,22 +15,11 @@
 #endif
 
 #define __HAVE_ARCH_MEMCPY 1
-#if defined(__SANITIZE_MEMORY__) && defined(__NO_FORTIFY)
-#undef memcpy
-#define memcpy __msan_memcpy
-#else
 extern void *memcpy(void *to, const void *from, size_t len);
-#endif
 extern void *__memcpy(void *to, const void *from, size_t len);
 
 #define __HAVE_ARCH_MEMSET
-#if defined(__SANITIZE_MEMORY__) && defined(__NO_FORTIFY)
-extern void *__msan_memset(void *s, int c, size_t n);
-#undef memset
-#define memset __msan_memset
-#else
 void *memset(void *s, int c, size_t n);
-#endif
 void *__memset(void *s, int c, size_t n);
 
 #define __HAVE_ARCH_MEMSET16
@@ -70,13 +59,7 @@ static inline void *memset64(uint64_t *s, uint64_t v, size_t n)
 }
 
 #define __HAVE_ARCH_MEMMOVE
-#if defined(__SANITIZE_MEMORY__) && defined(__NO_FORTIFY)
-#undef memmove
-void *__msan_memmove(void *dest, const void *src, size_t len);
-#define memmove __msan_memmove
-#else
 void *memmove(void *dest, const void *src, size_t count);
-#endif
 void *__memmove(void *dest, const void *src, size_t count);
 
 int memcmp(const void *cs, const void *ct, size_t count);
-- 
2.39.2.722.g9855ee24e9-goog

[PATCH 2/4] kmsan: another take at fixing memcpy tests

[Alexander Potapenko]

5478afc55a21 ("kmsan: fix memcpy tests") uses OPTIMIZER_HIDE_VAR() to
hide the uninitialized var from the compiler optimizations.

However OPTIMIZER_HIDE_VAR(uninit) enforces an immediate check of
@uninit, so memcpy tests did not actually check the behavior of memcpy(),
because they always contained a KMSAN report.

Replace OPTIMIZER_HIDE_VAR() with a file-local asm macro that just
clobbers the memory, and add a test case for memcpy() that does not
expect an error report.

Also reflow kmsan_test.c with clang-format.

Signed-off-by: Alexander Potapenko <glider@google.com>
---
 mm/kmsan/kmsan_test.c | 43 +++++++++++++++++++++++++++++++++++++------
 1 file changed, 37 insertions(+), 6 deletions(-)

diff --git a/mm/kmsan/kmsan_test.c b/mm/kmsan/kmsan_test.c
index 088e21a48dc4b..cc98a3f4e0899 100644
--- a/mm/kmsan/kmsan_test.c
+++ b/mm/kmsan/kmsan_test.c
@@ -407,6 +407,36 @@ static void test_printk(struct kunit *test)
 	KUNIT_EXPECT_TRUE(test, report_matches(&expect));
 }
 
+/*
+ * Prevent the compiler from optimizing @var away. Without this, Clang may
+ * notice that @var is uninitialized and drop memcpy() calls that use it.
+ *
+ * There is OPTIMIZER_HIDE_VAR() in linux/compier.h that we cannot use here,
+ * because it is implemented as inline assembly receiving @var as a parameter
+ * and will enforce a KMSAN check.
+ */
+#define DO_NOT_OPTIMIZE(var) asm("" ::: "memory")
+
+/*
+ * Test case: ensure that memcpy() correctly copies initialized values.
+ */
+static void test_init_memcpy(struct kunit *test)
+{
+	EXPECTATION_NO_REPORT(expect);
+	volatile int src;
+	volatile int dst = 0;
+
+	// Ensure DO_NOT_OPTIMIZE() does not cause extra checks.
+	DO_NOT_OPTIMIZE(src);
+	src = 1;
+	kunit_info(
+		test,
+		"memcpy()ing aligned initialized src to aligned dst (no reports)\n");
+	memcpy((void *)&dst, (void *)&src, sizeof(src));
+	kmsan_check_memory((void *)&dst, sizeof(dst));
+	KUNIT_EXPECT_TRUE(test, report_matches(&expect));
+}
+
 /*
  * Test case: ensure that memcpy() correctly copies uninitialized values between
  * aligned `src` and `dst`.
@@ -420,7 +450,7 @@ static void test_memcpy_aligned_to_aligned(struct kunit *test)
 	kunit_info(
 		test,
 		"memcpy()ing aligned uninit src to aligned dst (UMR report)\n");
-	OPTIMIZER_HIDE_VAR(uninit_src);
+	DO_NOT_OPTIMIZE(uninit_src);
 	memcpy((void *)&dst, (void *)&uninit_src, sizeof(uninit_src));
 	kmsan_check_memory((void *)&dst, sizeof(dst));
 	KUNIT_EXPECT_TRUE(test, report_matches(&expect));
@@ -443,7 +473,7 @@ static void test_memcpy_aligned_to_unaligned(struct kunit *test)
 	kunit_info(
 		test,
 		"memcpy()ing aligned uninit src to unaligned dst (UMR report)\n");
-	OPTIMIZER_HIDE_VAR(uninit_src);
+	DO_NOT_OPTIMIZE(uninit_src);
 	memcpy((void *)&dst[1], (void *)&uninit_src, sizeof(uninit_src));
 	kmsan_check_memory((void *)dst, 4);
 	KUNIT_EXPECT_TRUE(test, report_matches(&expect));
@@ -467,13 +497,14 @@ static void test_memcpy_aligned_to_unaligned2(struct kunit *test)
 	kunit_info(
 		test,
 		"memcpy()ing aligned uninit src to unaligned dst - part 2 (UMR report)\n");
-	OPTIMIZER_HIDE_VAR(uninit_src);
+	DO_NOT_OPTIMIZE(uninit_src);
 	memcpy((void *)&dst[1], (void *)&uninit_src, sizeof(uninit_src));
 	kmsan_check_memory((void *)&dst[4], sizeof(uninit_src));
 	KUNIT_EXPECT_TRUE(test, report_matches(&expect));
 }
 
-static noinline void fibonacci(int *array, int size, int start) {
+static noinline void fibonacci(int *array, int size, int start)
+{
 	if (start < 2 || (start == size))
 		return;
 	array[start] = array[start - 1] + array[start - 2];
@@ -482,8 +513,7 @@ static noinline void fibonacci(int *array, int size, int start) {
 
 static void test_long_origin_chain(struct kunit *test)
 {
-	EXPECTATION_UNINIT_VALUE_FN(expect,
-				    "test_long_origin_chain");
+	EXPECTATION_UNINIT_VALUE_FN(expect, "test_long_origin_chain");
 	/* (KMSAN_MAX_ORIGIN_DEPTH * 2) recursive calls to fibonacci(). */
 	volatile int accum[KMSAN_MAX_ORIGIN_DEPTH * 2 + 2];
 	int last = ARRAY_SIZE(accum) - 1;
@@ -515,6 +545,7 @@ static struct kunit_case kmsan_test_cases[] = {
 	KUNIT_CASE(test_uaf),
 	KUNIT_CASE(test_percpu_propagate),
 	KUNIT_CASE(test_printk),
+	KUNIT_CASE(test_init_memcpy),
 	KUNIT_CASE(test_memcpy_aligned_to_aligned),
 	KUNIT_CASE(test_memcpy_aligned_to_unaligned),
 	KUNIT_CASE(test_memcpy_aligned_to_unaligned2),
-- 
2.39.2.722.g9855ee24e9-goog

[PATCH 3/4] x86: kmsan: use C versions of memset16/memset32/memset64

[Alexander Potapenko]

KMSAN must see as many memory accesses as possible to prevent false
positive reports. Fall back to versions of
memset16()/memset32()/memset64() implemented in lib/string.c instead of
those written in assembly.

Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Helge Deller <deller@gmx.de>
Suggested-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: Alexander Potapenko <glider@google.com>
---
 arch/x86/include/asm/string_64.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/arch/x86/include/asm/string_64.h b/arch/x86/include/asm/string_64.h
index 9be401d971a99..e9c736f4686f5 100644
--- a/arch/x86/include/asm/string_64.h
+++ b/arch/x86/include/asm/string_64.h
@@ -22,6 +22,11 @@ extern void *__memcpy(void *to, const void *from, size_t len);
 void *memset(void *s, int c, size_t n);
 void *__memset(void *s, int c, size_t n);
 
+/*
+ * KMSAN needs to instrument as much code as possible. Use C versions of
+ * memsetXX() from lib/string.c under KMSAN.
+ */
+#if !defined(CONFIG_KMSAN)
 #define __HAVE_ARCH_MEMSET16
 static inline void *memset16(uint16_t *s, uint16_t v, size_t n)
 {
@@ -57,6 +62,7 @@ static inline void *memset64(uint64_t *s, uint64_t v, size_t n)
 		     : "memory");
 	return s;
 }
+#endif
 
 #define __HAVE_ARCH_MEMMOVE
 void *memmove(void *dest, const void *src, size_t count);
-- 
2.39.2.722.g9855ee24e9-goog

[PATCH 4/4] kmsan: add memsetXX tests

[Alexander Potapenko]

Add tests ensuring that memset16()/memset32()/memset64() are
instrumented by KMSAN and correctly initialize the memory.

Signed-off-by: Alexander Potapenko <glider@google.com>
---
 mm/kmsan/kmsan_test.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/mm/kmsan/kmsan_test.c b/mm/kmsan/kmsan_test.c
index cc98a3f4e0899..e450a000441fb 100644
--- a/mm/kmsan/kmsan_test.c
+++ b/mm/kmsan/kmsan_test.c
@@ -503,6 +503,25 @@ static void test_memcpy_aligned_to_unaligned2(struct kunit *test)
 	KUNIT_EXPECT_TRUE(test, report_matches(&expect));
 }
 
+/* Generate test cases for memset16(), memset32(), memset64(). */
+#define DEFINE_TEST_MEMSETXX(size, var_ty)                                  \
+	static void test_memset##size(struct kunit *test)                   \
+	{                                                                   \
+		EXPECTATION_NO_REPORT(expect);                              \
+		volatile var_ty uninit;                                     \
+                                                                            \
+		kunit_info(test,                                            \
+			   "memset" #size "() should initialize memory\n"); \
+		DO_NOT_OPTIMIZE(uninit);                                    \
+		memset##size((var_ty *)&uninit, 0, 1);                      \
+		kmsan_check_memory((void *)&uninit, sizeof(uninit));        \
+		KUNIT_EXPECT_TRUE(test, report_matches(&expect));           \
+	}
+
+DEFINE_TEST_MEMSETXX(16, uint16_t)
+DEFINE_TEST_MEMSETXX(32, uint32_t)
+DEFINE_TEST_MEMSETXX(64, uint64_t)
+
 static noinline void fibonacci(int *array, int size, int start)
 {
 	if (start < 2 || (start == size))
@@ -549,6 +568,9 @@ static struct kunit_case kmsan_test_cases[] = {
 	KUNIT_CASE(test_memcpy_aligned_to_aligned),
 	KUNIT_CASE(test_memcpy_aligned_to_unaligned),
 	KUNIT_CASE(test_memcpy_aligned_to_unaligned2),
+	KUNIT_CASE(test_memset16),
+	KUNIT_CASE(test_memset32),
+	KUNIT_CASE(test_memset64),
 	KUNIT_CASE(test_long_origin_chain),
 	{},
 };
-- 
2.39.2.722.g9855ee24e9-goog

Re: [PATCH 1/4] x86: kmsan: Don't rename memintrinsics in uninstrumented files

[Marco Elver]

On Wed, 1 Mar 2023 at 15:39, Alexander Potapenko <glider@google.com> wrote:
>
> KMSAN should be overriding calls to memset/memcpy/memmove and their

You mean that the compiler will override calls?
All supported compilers that have fsanitize=kernel-memory replace
memintrinsics with __msan_mem*() calls, right?

> __builtin_ versions in instrumented files, so there is no need to
> override them. In non-instrumented versions we are now required to
> leave memset() and friends intact, so we cannot replace them with
> __msan_XXX() functions.
>
> Cc: Kees Cook <keescook@chromium.org>
> Suggested-by: Marco Elver <elver@google.com>
> Signed-off-by: Alexander Potapenko <glider@google.com>

Other than that,

Reviewed-by: Marco Elver <elver@google.com>

> ---
>  arch/x86/include/asm/string_64.h | 17 -----------------
>  1 file changed, 17 deletions(-)
>
> diff --git a/arch/x86/include/asm/string_64.h b/arch/x86/include/asm/string_64.h
> index 888731ccf1f67..9be401d971a99 100644
> --- a/arch/x86/include/asm/string_64.h
> +++ b/arch/x86/include/asm/string_64.h
> @@ -15,22 +15,11 @@
>  #endif
>
>  #define __HAVE_ARCH_MEMCPY 1
> -#if defined(__SANITIZE_MEMORY__) && defined(__NO_FORTIFY)
> -#undef memcpy
> -#define memcpy __msan_memcpy
> -#else
>  extern void *memcpy(void *to, const void *from, size_t len);
> -#endif
>  extern void *__memcpy(void *to, const void *from, size_t len);
>
>  #define __HAVE_ARCH_MEMSET
> -#if defined(__SANITIZE_MEMORY__) && defined(__NO_FORTIFY)
> -extern void *__msan_memset(void *s, int c, size_t n);
> -#undef memset
> -#define memset __msan_memset
> -#else
>  void *memset(void *s, int c, size_t n);
> -#endif
>  void *__memset(void *s, int c, size_t n);
>
>  #define __HAVE_ARCH_MEMSET16
> @@ -70,13 +59,7 @@ static inline void *memset64(uint64_t *s, uint64_t v, size_t n)
>  }
>
>  #define __HAVE_ARCH_MEMMOVE
> -#if defined(__SANITIZE_MEMORY__) && defined(__NO_FORTIFY)
> -#undef memmove
> -void *__msan_memmove(void *dest, const void *src, size_t len);
> -#define memmove __msan_memmove
> -#else
>  void *memmove(void *dest, const void *src, size_t count);
> -#endif
>  void *__memmove(void *dest, const void *src, size_t count);
>
>  int memcmp(const void *cs, const void *ct, size_t count);
> --
> 2.39.2.722.g9855ee24e9-goog
>

Re: [PATCH 2/4] kmsan: another take at fixing memcpy tests

[Marco Elver]

On Wed, 1 Mar 2023 at 15:39, 'Alexander Potapenko' via kasan-dev
<kasan-dev@googlegroups.com> wrote:
>
> 5478afc55a21 ("kmsan: fix memcpy tests") uses OPTIMIZER_HIDE_VAR() to
> hide the uninitialized var from the compiler optimizations.
>
> However OPTIMIZER_HIDE_VAR(uninit) enforces an immediate check of
> @uninit, so memcpy tests did not actually check the behavior of memcpy(),
> because they always contained a KMSAN report.
>
> Replace OPTIMIZER_HIDE_VAR() with a file-local asm macro that just
> clobbers the memory, and add a test case for memcpy() that does not
> expect an error report.
>
> Also reflow kmsan_test.c with clang-format.
>
> Signed-off-by: Alexander Potapenko <glider@google.com>
> ---
>  mm/kmsan/kmsan_test.c | 43 +++++++++++++++++++++++++++++++++++++------
>  1 file changed, 37 insertions(+), 6 deletions(-)
>
> diff --git a/mm/kmsan/kmsan_test.c b/mm/kmsan/kmsan_test.c
> index 088e21a48dc4b..cc98a3f4e0899 100644
> --- a/mm/kmsan/kmsan_test.c
> +++ b/mm/kmsan/kmsan_test.c
> @@ -407,6 +407,36 @@ static void test_printk(struct kunit *test)
>         KUNIT_EXPECT_TRUE(test, report_matches(&expect));
>  }
>
> +/*
> + * Prevent the compiler from optimizing @var away. Without this, Clang may
> + * notice that @var is uninitialized and drop memcpy() calls that use it.
> + *
> + * There is OPTIMIZER_HIDE_VAR() in linux/compier.h that we cannot use here,
> + * because it is implemented as inline assembly receiving @var as a parameter
> + * and will enforce a KMSAN check.
> + */
> +#define DO_NOT_OPTIMIZE(var) asm("" ::: "memory")

That's just a normal "barrier()" - use that instead?

> +/*
> + * Test case: ensure that memcpy() correctly copies initialized values.
> + */
> +static void test_init_memcpy(struct kunit *test)
> +{
> +       EXPECTATION_NO_REPORT(expect);
> +       volatile int src;
> +       volatile int dst = 0;
> +
> +       // Ensure DO_NOT_OPTIMIZE() does not cause extra checks.

^^ this comment seems redundant now, given DO_NOT_OPTIMIZE() has a
comment (it's also using //-style comment).

> +       DO_NOT_OPTIMIZE(src);
> +       src = 1;
> +       kunit_info(
> +               test,
> +               "memcpy()ing aligned initialized src to aligned dst (no reports)\n");
> +       memcpy((void *)&dst, (void *)&src, sizeof(src));
> +       kmsan_check_memory((void *)&dst, sizeof(dst));
> +       KUNIT_EXPECT_TRUE(test, report_matches(&expect));
> +}
> +
>  /*
>   * Test case: ensure that memcpy() correctly copies uninitialized values between
>   * aligned `src` and `dst`.
> @@ -420,7 +450,7 @@ static void test_memcpy_aligned_to_aligned(struct kunit *test)
>         kunit_info(
>                 test,
>                 "memcpy()ing aligned uninit src to aligned dst (UMR report)\n");
> -       OPTIMIZER_HIDE_VAR(uninit_src);
> +       DO_NOT_OPTIMIZE(uninit_src);
>         memcpy((void *)&dst, (void *)&uninit_src, sizeof(uninit_src));
>         kmsan_check_memory((void *)&dst, sizeof(dst));
>         KUNIT_EXPECT_TRUE(test, report_matches(&expect));
> @@ -443,7 +473,7 @@ static void test_memcpy_aligned_to_unaligned(struct kunit *test)
>         kunit_info(
>                 test,
>                 "memcpy()ing aligned uninit src to unaligned dst (UMR report)\n");
> -       OPTIMIZER_HIDE_VAR(uninit_src);
> +       DO_NOT_OPTIMIZE(uninit_src);
>         memcpy((void *)&dst[1], (void *)&uninit_src, sizeof(uninit_src));
>         kmsan_check_memory((void *)dst, 4);
>         KUNIT_EXPECT_TRUE(test, report_matches(&expect));
> @@ -467,13 +497,14 @@ static void test_memcpy_aligned_to_unaligned2(struct kunit *test)
>         kunit_info(
>                 test,
>                 "memcpy()ing aligned uninit src to unaligned dst - part 2 (UMR report)\n");
> -       OPTIMIZER_HIDE_VAR(uninit_src);
> +       DO_NOT_OPTIMIZE(uninit_src);
>         memcpy((void *)&dst[1], (void *)&uninit_src, sizeof(uninit_src));
>         kmsan_check_memory((void *)&dst[4], sizeof(uninit_src));
>         KUNIT_EXPECT_TRUE(test, report_matches(&expect));
>  }
>
> -static noinline void fibonacci(int *array, int size, int start) {
> +static noinline void fibonacci(int *array, int size, int start)
> +{
>         if (start < 2 || (start == size))
>                 return;
>         array[start] = array[start - 1] + array[start - 2];
> @@ -482,8 +513,7 @@ static noinline void fibonacci(int *array, int size, int start) {
>
>  static void test_long_origin_chain(struct kunit *test)
>  {
> -       EXPECTATION_UNINIT_VALUE_FN(expect,
> -                                   "test_long_origin_chain");
> +       EXPECTATION_UNINIT_VALUE_FN(expect, "test_long_origin_chain");
>         /* (KMSAN_MAX_ORIGIN_DEPTH * 2) recursive calls to fibonacci(). */
>         volatile int accum[KMSAN_MAX_ORIGIN_DEPTH * 2 + 2];
>         int last = ARRAY_SIZE(accum) - 1;
> @@ -515,6 +545,7 @@ static struct kunit_case kmsan_test_cases[] = {
>         KUNIT_CASE(test_uaf),
>         KUNIT_CASE(test_percpu_propagate),
>         KUNIT_CASE(test_printk),
> +       KUNIT_CASE(test_init_memcpy),
>         KUNIT_CASE(test_memcpy_aligned_to_aligned),
>         KUNIT_CASE(test_memcpy_aligned_to_unaligned),
>         KUNIT_CASE(test_memcpy_aligned_to_unaligned2),
> --
> 2.39.2.722.g9855ee24e9-goog
>
> --
> 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+unsubscribe@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/kasan-dev/20230301143933.2374658-2-glider%40google.com.

Re: [PATCH 3/4] x86: kmsan: use C versions of memset16/memset32/memset64

[Marco Elver]

On Wed, 1 Mar 2023 at 15:39, 'Alexander Potapenko' via kasan-dev
<kasan-dev@googlegroups.com> wrote:
>
> KMSAN must see as many memory accesses as possible to prevent false
> positive reports. Fall back to versions of
> memset16()/memset32()/memset64() implemented in lib/string.c instead of
> those written in assembly.
>
> Cc: Geert Uytterhoeven <geert@linux-m68k.org>
> Cc: Daniel Vetter <daniel@ffwll.ch>
> Cc: Helge Deller <deller@gmx.de>
> Suggested-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
> Signed-off-by: Alexander Potapenko <glider@google.com>

Reviewed-by: Marco Elver <elver@google.com>

> ---
>  arch/x86/include/asm/string_64.h | 6 ++++++
>  1 file changed, 6 insertions(+)
>
> diff --git a/arch/x86/include/asm/string_64.h b/arch/x86/include/asm/string_64.h
> index 9be401d971a99..e9c736f4686f5 100644
> --- a/arch/x86/include/asm/string_64.h
> +++ b/arch/x86/include/asm/string_64.h
> @@ -22,6 +22,11 @@ extern void *__memcpy(void *to, const void *from, size_t len);
>  void *memset(void *s, int c, size_t n);
>  void *__memset(void *s, int c, size_t n);
>
> +/*
> + * KMSAN needs to instrument as much code as possible. Use C versions of
> + * memsetXX() from lib/string.c under KMSAN.
> + */
> +#if !defined(CONFIG_KMSAN)
>  #define __HAVE_ARCH_MEMSET16
>  static inline void *memset16(uint16_t *s, uint16_t v, size_t n)
>  {
> @@ -57,6 +62,7 @@ static inline void *memset64(uint64_t *s, uint64_t v, size_t n)
>                      : "memory");
>         return s;
>  }
> +#endif
>
>  #define __HAVE_ARCH_MEMMOVE
>  void *memmove(void *dest, const void *src, size_t count);
> --
> 2.39.2.722.g9855ee24e9-goog
>
> --
> 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+unsubscribe@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/kasan-dev/20230301143933.2374658-3-glider%40google.com.

Re: [PATCH 4/4] kmsan: add memsetXX tests

[Marco Elver]

On Wed, 1 Mar 2023 at 15:39, Alexander Potapenko <glider@google.com> wrote:
>
> Add tests ensuring that memset16()/memset32()/memset64() are
> instrumented by KMSAN and correctly initialize the memory.
>
> Signed-off-by: Alexander Potapenko <glider@google.com>
> ---
>  mm/kmsan/kmsan_test.c | 22 ++++++++++++++++++++++
>  1 file changed, 22 insertions(+)
>
> diff --git a/mm/kmsan/kmsan_test.c b/mm/kmsan/kmsan_test.c
> index cc98a3f4e0899..e450a000441fb 100644
> --- a/mm/kmsan/kmsan_test.c
> +++ b/mm/kmsan/kmsan_test.c
> @@ -503,6 +503,25 @@ static void test_memcpy_aligned_to_unaligned2(struct kunit *test)
>         KUNIT_EXPECT_TRUE(test, report_matches(&expect));
>  }
>
> +/* Generate test cases for memset16(), memset32(), memset64(). */
> +#define DEFINE_TEST_MEMSETXX(size, var_ty)                                  \
> +       static void test_memset##size(struct kunit *test)                   \
> +       {                                                                   \
> +               EXPECTATION_NO_REPORT(expect);                              \
> +               volatile var_ty uninit;                                     \

This could just be 'uint##size##_t' and you can drop 'var_ty'.

> +                                                                            \
> +               kunit_info(test,                                            \
> +                          "memset" #size "() should initialize memory\n"); \
> +               DO_NOT_OPTIMIZE(uninit);                                    \
> +               memset##size((var_ty *)&uninit, 0, 1);                      \
> +               kmsan_check_memory((void *)&uninit, sizeof(uninit));        \
> +               KUNIT_EXPECT_TRUE(test, report_matches(&expect));           \
> +       }
> +
> +DEFINE_TEST_MEMSETXX(16, uint16_t)
> +DEFINE_TEST_MEMSETXX(32, uint32_t)
> +DEFINE_TEST_MEMSETXX(64, uint64_t)
> +
>  static noinline void fibonacci(int *array, int size, int start)
>  {
>         if (start < 2 || (start == size))
> @@ -549,6 +568,9 @@ static struct kunit_case kmsan_test_cases[] = {
>         KUNIT_CASE(test_memcpy_aligned_to_aligned),
>         KUNIT_CASE(test_memcpy_aligned_to_unaligned),
>         KUNIT_CASE(test_memcpy_aligned_to_unaligned2),
> +       KUNIT_CASE(test_memset16),
> +       KUNIT_CASE(test_memset32),
> +       KUNIT_CASE(test_memset64),
>         KUNIT_CASE(test_long_origin_chain),
>         {},
>  };
> --
> 2.39.2.722.g9855ee24e9-goog
>

Re: [PATCH 4/4] kmsan: add memsetXX tests

[Alexander Potapenko]

On Thu, Mar 2, 2023 at 12:23 PM Marco Elver <elver@google.com> wrote:
>
> On Wed, 1 Mar 2023 at 15:39, Alexander Potapenko <glider@google.com> wrote:
> >
> > Add tests ensuring that memset16()/memset32()/memset64() are
> > instrumented by KMSAN and correctly initialize the memory.
> >
> > Signed-off-by: Alexander Potapenko <glider@google.com>
> > ---
> >  mm/kmsan/kmsan_test.c | 22 ++++++++++++++++++++++
> >  1 file changed, 22 insertions(+)
> >
> > diff --git a/mm/kmsan/kmsan_test.c b/mm/kmsan/kmsan_test.c
> > index cc98a3f4e0899..e450a000441fb 100644
> > --- a/mm/kmsan/kmsan_test.c
> > +++ b/mm/kmsan/kmsan_test.c
> > @@ -503,6 +503,25 @@ static void test_memcpy_aligned_to_unaligned2(struct kunit *test)
> >         KUNIT_EXPECT_TRUE(test, report_matches(&expect));
> >  }
> >
> > +/* Generate test cases for memset16(), memset32(), memset64(). */
> > +#define DEFINE_TEST_MEMSETXX(size, var_ty)                                  \
> > +       static void test_memset##size(struct kunit *test)                   \
> > +       {                                                                   \
> > +               EXPECTATION_NO_REPORT(expect);                              \
> > +               volatile var_ty uninit;                                     \
>
> This could just be 'uint##size##_t' and you can drop 'var_ty'.

Indeed, thanks!

Re: [PATCH 1/4] x86: kmsan: Don't rename memintrinsics in uninstrumented files

[Alexander Potapenko]

On Thu, Mar 2, 2023 at 12:14 PM Marco Elver <elver@google.com> wrote:
>
> On Wed, 1 Mar 2023 at 15:39, Alexander Potapenko <glider@google.com> wrote:
> >
> > KMSAN should be overriding calls to memset/memcpy/memmove and their
>
> You mean that the compiler will override calls?
> All supported compilers that have fsanitize=kernel-memory replace
> memintrinsics with __msan_mem*() calls, right?

Right. Changed to:

KMSAN already replaces calls to to memset/memcpy/memmove and their
__builtin_ versions with __msan_memset/__msan_memcpy/__msan_memmove in
instrumented files, so there is no need to override them.


>
> > __builtin_ versions in instrumented files, so there is no need to
> > override them. In non-instrumented versions we are now required to
> > leave memset() and friends intact, so we cannot replace them with
> > __msan_XXX() functions.
> >
> > Cc: Kees Cook <keescook@chromium.org>
> > Suggested-by: Marco Elver <elver@google.com>
> > Signed-off-by: Alexander Potapenko <glider@google.com>
>
> Other than that,
>
> Reviewed-by: Marco Elver <elver@google.com>
Thanks!

Re: [PATCH 2/4] kmsan: another take at fixing memcpy tests

[Alexander Potapenko]

> > +#define DO_NOT_OPTIMIZE(var) asm("" ::: "memory")
>
> That's just a normal "barrier()" - use that instead?

Ok, will do (I still think I'd better hide it behind a macro so that
we can change the implementation of DO_NOT_OPTIMIZE in the future if
the compiler starts outsmarting us again.

> > +/*
> > + * Test case: ensure that memcpy() correctly copies initialized values.
> > + */
> > +static void test_init_memcpy(struct kunit *test)
> > +{
> > +       EXPECTATION_NO_REPORT(expect);
> > +       volatile int src;
> > +       volatile int dst = 0;
> > +
> > +       // Ensure DO_NOT_OPTIMIZE() does not cause extra checks.
>
> ^^ this comment seems redundant now, given DO_NOT_OPTIMIZE() has a
> comment (it's also using //-style comment).

Moved it to the test description:

/*
 * Test case: ensure that memcpy() correctly copies initialized values.
 * Also serves as a regression test to ensure DO_NOT_OPTIMIZE() does not cause
 * extra checks.
 */

I think it's still relevant here.

Re: [PATCH 1/4] x86: kmsan: Don't rename memintrinsics in uninstrumented files

[Marco Elver]

On Thu, 2 Mar 2023 at 15:28, Alexander Potapenko <glider@google.com> wrote:
>
> On Thu, Mar 2, 2023 at 12:14 PM Marco Elver <elver@google.com> wrote:
> >
> > On Wed, 1 Mar 2023 at 15:39, Alexander Potapenko <glider@google.com> wrote:
> > >
> > > KMSAN should be overriding calls to memset/memcpy/memmove and their
> >
> > You mean that the compiler will override calls?
> > All supported compilers that have fsanitize=kernel-memory replace
> > memintrinsics with __msan_mem*() calls, right?
>
> Right. Changed to:
>
> KMSAN already replaces calls to to memset/memcpy/memmove and their
> __builtin_ versions with __msan_memset/__msan_memcpy/__msan_memmove in
> instrumented files, so there is no need to override them.

But it's not KMSAN - KMSAN is the combined end result of runtime and
compiler - in this case we need to be specific and point out it's the
compiler that's doing it. There is no code in the Linux kernel that
does this replacement.

>
> >
> > > __builtin_ versions in instrumented files, so there is no need to
> > > override them. In non-instrumented versions we are now required to
> > > leave memset() and friends intact, so we cannot replace them with
> > > __msan_XXX() functions.
> > >
> > > Cc: Kees Cook <keescook@chromium.org>
> > > Suggested-by: Marco Elver <elver@google.com>
> > > Signed-off-by: Alexander Potapenko <glider@google.com>
> >
> > Other than that,
> >
> > Reviewed-by: Marco Elver <elver@google.com>
> Thanks!

Re: [PATCH 1/4] x86: kmsan: Don't rename memintrinsics in uninstrumented files

[Alexander Potapenko]

On Thu, Mar 2, 2023 at 4:13 PM Marco Elver <elver@google.com> wrote:
>
> On Thu, 2 Mar 2023 at 15:28, Alexander Potapenko <glider@google.com> wrote:
> >
> > On Thu, Mar 2, 2023 at 12:14 PM Marco Elver <elver@google.com> wrote:
> > >
> > > On Wed, 1 Mar 2023 at 15:39, Alexander Potapenko <glider@google.com> wrote:
> > > >
> > > > KMSAN should be overriding calls to memset/memcpy/memmove and their
> > >
> > > You mean that the compiler will override calls?
> > > All supported compilers that have fsanitize=kernel-memory replace
> > > memintrinsics with __msan_mem*() calls, right?
> >
> > Right. Changed to:
> >
> > KMSAN already replaces calls to to memset/memcpy/memmove and their
> > __builtin_ versions with __msan_memset/__msan_memcpy/__msan_memmove in
> > instrumented files, so there is no need to override them.
>
> But it's not KMSAN - KMSAN is the combined end result of runtime and
> compiler - in this case we need to be specific and point out it's the
> compiler that's doing it. There is no code in the Linux kernel that
> does this replacement.

Agreed. I'll replace with "clang -fsanitize=kernel-memory"

Re: [PATCH 3/4] x86: kmsan: use C versions of memset16/memset32/memset64

[Alexander Potapenko]

This is the second version of the patch. Sorry for the inconvenience.

[PATCH 3/4] x86: kmsan: use C versions of memset16/memset32/memset64

[Alexander Potapenko]

KMSAN must see as many memory accesses as possible to prevent false
positive reports. Fall back to versions of
memset16()/memset32()/memset64() implemented in lib/string.c instead of
those written in assembly.

Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Helge Deller <deller@gmx.de>
Suggested-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: Alexander Potapenko <glider@google.com>
Reviewed-by: Marco Elver <elver@google.com>
---
 arch/x86/include/asm/string_64.h | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/arch/x86/include/asm/string_64.h b/arch/x86/include/asm/string_64.h
index 9be401d971a99..e9c736f4686f5 100644
--- a/arch/x86/include/asm/string_64.h
+++ b/arch/x86/include/asm/string_64.h
@@ -22,6 +22,11 @@ extern void *__memcpy(void *to, const void *from, size_t len);
 void *memset(void *s, int c, size_t n);
 void *__memset(void *s, int c, size_t n);
 
+/*
+ * KMSAN needs to instrument as much code as possible. Use C versions of
+ * memsetXX() from lib/string.c under KMSAN.
+ */
+#if !defined(CONFIG_KMSAN)
 #define __HAVE_ARCH_MEMSET16
 static inline void *memset16(uint16_t *s, uint16_t v, size_t n)
 {
@@ -57,6 +62,7 @@ static inline void *memset64(uint64_t *s, uint64_t v, size_t n)
 		     : "memory");
 	return s;
 }
+#endif
 
 #define __HAVE_ARCH_MEMMOVE
 void *memmove(void *dest, const void *src, size_t count);
-- 
2.40.0.rc0.216.gc4246ad0f0-goog