Commit 7771bdbb authored by Andrey Ryabinin's avatar Andrey Ryabinin Committed by Linus Torvalds
Browse files

kasan: remove use after scope bugs detection.

Use after scope bugs detector seems to be almost entirely useless for
the linux kernel.  It exists over two years, but I've seen only one
valid bug so far [1].  And the bug was fixed before it has been
reported.  There were some other use-after-scope reports, but they were
false-positives due to different reasons like incompatibility with
structleak plugin.

This feature significantly increases stack usage, especially with GCC <
9 version, and causes a 32K stack overflow.  It probably adds
performance penalty too.

Given all that, let's remove use-after-scope detector entirely.

While preparing this patch I've noticed that we mistakenly enable
use-after-scope detection for clang compiler regardless of
CONFIG_KASAN_EXTRA setting.  This is also fixed now.

[1] http://lkml.kernel.org/r/<20171129052106.rhgbjhhis53hkgfn@wfg-t540p.sh.intel.com>

Link: http://lkml.kernel.org/r/20190111185842.13978-1-aryabinin@virtuozzo.com

Signed-off-by: default avatarAndrey Ryabinin <aryabinin@virtuozzo.com>
Acked-by: Will Deacon <will.deacon@arm.com>		[arm64]
Cc: Qian Cai <cai@lca.pw>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent 46612b75
......@@ -80,11 +80,7 @@
*/
#ifdef CONFIG_KASAN
#define KASAN_SHADOW_SIZE (UL(1) << (VA_BITS - KASAN_SHADOW_SCALE_SHIFT))
#ifdef CONFIG_KASAN_EXTRA
#define KASAN_THREAD_SHIFT 2
#else
#define KASAN_THREAD_SHIFT 1
#endif /* CONFIG_KASAN_EXTRA */
#else
#define KASAN_SHADOW_SIZE (0)
#define KASAN_THREAD_SHIFT 0
......
......@@ -222,7 +222,6 @@ config ENABLE_MUST_CHECK
config FRAME_WARN
int "Warn for stack frames larger than (needs gcc 4.4)"
range 0 8192
default 3072 if KASAN_EXTRA
default 2048 if GCC_PLUGIN_LATENT_ENTROPY
default 1280 if (!64BIT && PARISC)
default 1024 if (!64BIT && !PARISC)
......
......@@ -78,16 +78,6 @@ config KASAN_SW_TAGS
endchoice
config KASAN_EXTRA
bool "KASAN: extra checks"
depends on KASAN_GENERIC && DEBUG_KERNEL && !COMPILE_TEST
help
This enables further checks in generic KASAN, for now it only
includes the address-use-after-scope check that can lead to
excessive kernel stack usage, frame size warnings and longer
compile time.
See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81715
choice
prompt "Instrumentation type"
depends on KASAN
......
......@@ -480,29 +480,6 @@ static noinline void __init copy_user_test(void)
kfree(kmem);
}
static noinline void __init use_after_scope_test(void)
{
volatile char *volatile p;
pr_info("use-after-scope on int\n");
{
int local = 0;
p = (char *)&local;
}
p[0] = 1;
p[3] = 1;
pr_info("use-after-scope on array\n");
{
char local[1024] = {0};
p = local;
}
p[0] = 1;
p[1023] = 1;
}
static noinline void __init kasan_alloca_oob_left(void)
{
volatile int i = 10;
......@@ -682,7 +659,6 @@ static int __init kmalloc_tests_init(void)
kasan_alloca_oob_right();
ksize_unpoisons_memory();
copy_user_test();
use_after_scope_test();
kmem_cache_double_free();
kmem_cache_invalid_free();
kasan_memchr();
......
......@@ -275,25 +275,6 @@ EXPORT_SYMBOL(__asan_storeN_noabort);
void __asan_handle_no_return(void) {}
EXPORT_SYMBOL(__asan_handle_no_return);
/* Emitted by compiler to poison large objects when they go out of scope. */
void __asan_poison_stack_memory(const void *addr, size_t size)
{
/*
* Addr is KASAN_SHADOW_SCALE_SIZE-aligned and the object is surrounded
* by redzones, so we simply round up size to simplify logic.
*/
kasan_poison_shadow(addr, round_up(size, KASAN_SHADOW_SCALE_SIZE),
KASAN_USE_AFTER_SCOPE);
}
EXPORT_SYMBOL(__asan_poison_stack_memory);
/* Emitted by compiler to unpoison large objects when they go into scope. */
void __asan_unpoison_stack_memory(const void *addr, size_t size)
{
kasan_unpoison_shadow(addr, size);
}
EXPORT_SYMBOL(__asan_unpoison_stack_memory);
/* Emitted by compiler to poison alloca()ed objects. */
void __asan_alloca_poison(unsigned long addr, size_t size)
{
......
......@@ -82,9 +82,6 @@ static const char *get_shadow_bug_type(struct kasan_access_info *info)
case KASAN_KMALLOC_FREE:
bug_type = "use-after-free";
break;
case KASAN_USE_AFTER_SCOPE:
bug_type = "use-after-scope";
break;
case KASAN_ALLOCA_LEFT:
case KASAN_ALLOCA_RIGHT:
bug_type = "alloca-out-of-bounds";
......
......@@ -34,7 +34,6 @@
#define KASAN_STACK_MID 0xF2
#define KASAN_STACK_RIGHT 0xF3
#define KASAN_STACK_PARTIAL 0xF4
#define KASAN_USE_AFTER_SCOPE 0xF8
/*
* alloca redzone shadow values
......@@ -187,8 +186,6 @@ void __asan_unregister_globals(struct kasan_global *globals, size_t size);
void __asan_loadN(unsigned long addr, size_t size);
void __asan_storeN(unsigned long addr, size_t size);
void __asan_handle_no_return(void);
void __asan_poison_stack_memory(const void *addr, size_t size);
void __asan_unpoison_stack_memory(const void *addr, size_t size);
void __asan_alloca_poison(unsigned long addr, size_t size);
void __asan_allocas_unpoison(const void *stack_top, const void *stack_bottom);
......
......@@ -27,14 +27,9 @@ else
$(call cc-param,asan-globals=1) \
$(call cc-param,asan-instrumentation-with-call-threshold=$(call_threshold)) \
$(call cc-param,asan-stack=$(CONFIG_KASAN_STACK)) \
$(call cc-param,asan-use-after-scope=1) \
$(call cc-param,asan-instrument-allocas=1)
endif
ifdef CONFIG_KASAN_EXTRA
CFLAGS_KASAN += $(call cc-option, -fsanitize-address-use-after-scope)
endif
endif # CONFIG_KASAN_GENERIC
ifdef CONFIG_KASAN_SW_TAGS
......
......@@ -68,10 +68,6 @@ config GCC_PLUGIN_LATENT_ENTROPY
config GCC_PLUGIN_STRUCTLEAK
bool "Force initialization of variables containing userspace addresses"
# Currently STRUCTLEAK inserts initialization out of live scope of
# variables from KASAN point of view. This leads to KASAN false
# positive reports. Prohibit this combination for now.
depends on !KASAN_EXTRA
help
This plugin zero-initializes any structures containing a
__user attribute. This can prevent some classes of information
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment