[PATCH v2 1/2] kasan: detect negative size in memory operation function

[Walter Wu]

KASAN missed detecting size is negative numbers in memset(), memcpy(),
and memmove(), it will cause out-of-bounds bug, so needs to be detected
by KASAN.

If size is negative numbers, then it has three reasons to be
defined as heap-out-of-bounds bug type.
1) Casting negative numbers to size_t would indeed turn up as
   a large size_t and its value will be larger than ULONG_MAX/2,
   so that this can qualify as out-of-bounds.
2) If KASAN has new bug type and user-space passes negative size,
   then there are duplicate reports. So don't produce new bug type
   in order to prevent duplicate reports by some systems (e.g. syzbot)
   to report the same bug twice.
3) When size is negative numbers, it may be passed from user-space.
   So we always print heap-out-of-bounds in order to prevent that
   kernel-space and user-space have the same bug but have duplicate
   reports.

KASAN report:

 BUG: KASAN: heap-out-of-bounds in kmalloc_memmove_invalid_size+0x70/0xa0
 Read of size 18446744073709551608 at addr ffffff8069660904 by task cat/72

 CPU: 2 PID: 72 Comm: cat Not tainted 5.4.0-rc1-next-20191004ajb-00001-gdb8af2f372b2-dirty #1
 Hardware name: linux,dummy-virt (DT)
 Call trace:
  dump_backtrace+0x0/0x288
  show_stack+0x14/0x20
  dump_stack+0x10c/0x164
  print_address_description.isra.9+0x68/0x378
  __kasan_report+0x164/0x1a0
  kasan_report+0xc/0x18
  check_memory_region+0x174/0x1d0
  memmove+0x34/0x88
  kmalloc_memmove_invalid_size+0x70/0xa0

[1] https://bugzilla.kernel.org/show_bug.cgi?id=199341

Changes in v2:
Fix the indentation bug, thanks for the reminder Matthew.

Signed-off-by: Walter Wu <walter-zh.wu@xxxxxxxxxxxx>
Reported-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
Suggested-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
Reviewed-by: Dmitry Vyukov <dvyukov@xxxxxxxxxx>
Cc: Andrey Ryabinin <aryabinin@xxxxxxxxxxxxx>
Cc: Alexander Potapenko <glider@xxxxxxxxxx>
---
 mm/kasan/common.c         | 13 ++++++++-----
 mm/kasan/generic.c        |  5 +++++
 mm/kasan/generic_report.c | 18 ++++++++++++++++++
 mm/kasan/tags.c           |  5 +++++
 mm/kasan/tags_report.c    | 18 ++++++++++++++++++
 5 files changed, 54 insertions(+), 5 deletions(-)

diff --git a/mm/kasan/common.c b/mm/kasan/common.c
index 6814d6d6a023..16a370023425 100644
--- a/mm/kasan/common.c
+++ b/mm/kasan/common.c
@@ -102,7 +102,8 @@ EXPORT_SYMBOL(__kasan_check_write);
 #undef memset
 void *memset(void *addr, int c, size_t len)
 {
-	check_memory_region((unsigned long)addr, len, true, _RET_IP_);
+	if (!check_memory_region((unsigned long)addr, len, true, _RET_IP_))
+		return NULL;
 
 	return __memset(addr, c, len);
 }
@@ -110,8 +111,9 @@ void *memset(void *addr, int c, size_t len)
 #undef memmove
 void *memmove(void *dest, const void *src, size_t len)
 {
-	check_memory_region((unsigned long)src, len, false, _RET_IP_);
-	check_memory_region((unsigned long)dest, len, true, _RET_IP_);
+	if (!check_memory_region((unsigned long)src, len, false, _RET_IP_) ||
+	    !check_memory_region((unsigned long)dest, len, true, _RET_IP_))
+		return NULL;
 
 	return __memmove(dest, src, len);
 }
@@ -119,8 +121,9 @@ void *memmove(void *dest, const void *src, size_t len)
 #undef memcpy
 void *memcpy(void *dest, const void *src, size_t len)
 {
-	check_memory_region((unsigned long)src, len, false, _RET_IP_);
-	check_memory_region((unsigned long)dest, len, true, _RET_IP_);
+	if (!check_memory_region((unsigned long)src, len, false, _RET_IP_) ||
+	    !check_memory_region((unsigned long)dest, len, true, _RET_IP_))
+		return NULL;
 
 	return __memcpy(dest, src, len);
 }
diff --git a/mm/kasan/generic.c b/mm/kasan/generic.c
index 616f9dd82d12..02148a317d27 100644
--- a/mm/kasan/generic.c
+++ b/mm/kasan/generic.c
@@ -173,6 +173,11 @@ static __always_inline bool check_memory_region_inline(unsigned long addr,
 	if (unlikely(size == 0))
 		return true;
 
+	if (unlikely((long)size < 0)) {
+		kasan_report(addr, size, write, ret_ip);
+		return false;
+	}
+
 	if (unlikely((void *)addr <
 		kasan_shadow_to_mem((void *)KASAN_SHADOW_START))) {
 		kasan_report(addr, size, write, ret_ip);
diff --git a/mm/kasan/generic_report.c b/mm/kasan/generic_report.c
index 36c645939bc9..52a92c7db697 100644
--- a/mm/kasan/generic_report.c
+++ b/mm/kasan/generic_report.c
@@ -107,6 +107,24 @@ static const char *get_wild_bug_type(struct kasan_access_info *info)
 
 const char *get_bug_type(struct kasan_access_info *info)
 {
+	/*
+	 * If access_size is negative numbers, then it has three reasons
+	 * to be defined as heap-out-of-bounds bug type.
+	 * 1) Casting negative numbers to size_t would indeed turn up as
+	 *    a large size_t and its value will be larger than ULONG_MAX/2,
+	 *    so that this can qualify as out-of-bounds.
+	 * 2) If KASAN has new bug type and user-space passes negative size,
+	 *    then there are duplicate reports. So don't produce new bug type
+	 *    in order to prevent duplicate reports by some systems
+	 *    (e.g. syzbot) to report the same bug twice.
+	 * 3) When size is negative numbers, it may be passed from user-space.
+	 *    So we always print heap-out-of-bounds in order to prevent that
+	 *    kernel-space and user-space have the same bug but have duplicate
+	 *    reports.
+	 */
+	if ((long)info->access_size < 0)
+		return "heap-out-of-bounds";
+
 	if (addr_has_shadow(info->access_addr))
 		return get_shadow_bug_type(info);
 	return get_wild_bug_type(info);
diff --git a/mm/kasan/tags.c b/mm/kasan/tags.c
index 0e987c9ca052..b829535a3ad7 100644
--- a/mm/kasan/tags.c
+++ b/mm/kasan/tags.c
@@ -86,6 +86,11 @@ bool check_memory_region(unsigned long addr, size_t size, bool write,
 	if (unlikely(size == 0))
 		return true;
 
+	if (unlikely((long)size < 0)) {
+		kasan_report(addr, size, write, ret_ip);
+		return false;
+	}
+
 	tag = get_tag((const void *)addr);
 
 	/*
diff --git a/mm/kasan/tags_report.c b/mm/kasan/tags_report.c
index 969ae08f59d7..f7ae474aef3a 100644
--- a/mm/kasan/tags_report.c
+++ b/mm/kasan/tags_report.c
@@ -36,6 +36,24 @@
 
 const char *get_bug_type(struct kasan_access_info *info)
 {
+	/*
+	 * If access_size is negative numbers, then it has three reasons
+	 * to be defined as heap-out-of-bounds bug type.
+	 * 1) Casting negative numbers to size_t would indeed turn up as
+	 *    a large size_t and its value will be larger than ULONG_MAX/2,
+	 *    so that this can qualify as out-of-bounds.
+	 * 2) If KASAN has new bug type and user-space passes negative size,
+	 *    then there are duplicate reports. So don't produce new bug type
+	 *    in order to prevent duplicate reports by some systems
+	 *    (e.g. syzbot) to report the same bug twice.
+	 * 3) When size is negative numbers, it may be passed from user-space.
+	 *    So we always print heap-out-of-bounds in order to prevent that
+	 *    kernel-space and user-space have the same bug but have duplicate
+	 *    reports.
+	 */
+	if ((long)info->access_size < 0)
+		return "heap-out-of-bounds";
+
 #ifdef CONFIG_KASAN_SW_TAGS_IDENTIFY
 	struct kasan_alloc_meta *alloc_meta;
 	struct kmem_cache *cache;
-- 
2.18.0