[PATCH] kasan: remove redundant config option

[Walter Wu]

CONFIG_KASAN_STACK and CONFIG_KASAN_STACK_ENABLE both enable KASAN
stack instrumentation, but we should only need one config option,
so that we remove CONFIG_KASAN_STACK_ENABLE. see [1].

For gcc we could do no prompt and default value y, and for clang
prompt and default value n.

[1]: https://bugzilla.kernel.org/show_bug.cgi?id=210221

Signed-off-by: Walter Wu <walter...@mediatek.com>
Suggested-by: Dmitry Vyukov <dvy...@google.com>
Cc: Andrey Ryabinin <arya...@virtuozzo.com>
Cc: Dmitry Vyukov <dvy...@google.com>
Cc: Andrey Konovalov <andre...@google.com>
Cc: Alexander Potapenko <gli...@google.com>
Cc: Andrew Morton <ak...@linux-foundation.org>
---
arch/arm64/kernel/sleep.S | 2 +-
arch/x86/kernel/acpi/wakeup_64.S | 2 +-
include/linux/kasan.h | 2 +-
lib/Kconfig.kasan | 11 ++++-------
mm/kasan/common.c | 2 +-
mm/kasan/kasan.h | 2 +-
mm/kasan/report_generic.c | 2 +-
scripts/Makefile.kasan | 10 ++++++++--
8 files changed, 18 insertions(+), 15 deletions(-)

diff --git a/arch/arm64/kernel/sleep.S b/arch/arm64/kernel/sleep.S
index 6bdef7362c0e..7c44ede122a9 100644
--- a/arch/arm64/kernel/sleep.S
+++ b/arch/arm64/kernel/sleep.S
@@ -133,7 +133,7 @@ SYM_FUNC_START(_cpu_resume)
*/
bl cpu_do_resume

-#if defined(CONFIG_KASAN) && CONFIG_KASAN_STACK
+#if defined(CONFIG_KASAN) && defined(CONFIG_KASAN_STACK)
mov x0, sp
bl kasan_unpoison_task_stack_below
#endif
diff --git a/arch/x86/kernel/acpi/wakeup_64.S b/arch/x86/kernel/acpi/wakeup_64.S
index 5d3a0b8fd379..c7f412f4e07d 100644
--- a/arch/x86/kernel/acpi/wakeup_64.S
+++ b/arch/x86/kernel/acpi/wakeup_64.S
@@ -112,7 +112,7 @@ SYM_FUNC_START(do_suspend_lowlevel)
movq pt_regs_r14(%rax), %r14
movq pt_regs_r15(%rax), %r15

-#if defined(CONFIG_KASAN) && CONFIG_KASAN_STACK
+#if defined(CONFIG_KASAN) && defined(CONFIG_KASAN_STACK)
/*
* The suspend path may have poisoned some areas deeper in the stack,
* which we now need to unpoison.
diff --git a/include/linux/kasan.h b/include/linux/kasan.h
index 5e0655fb2a6f..35d1e9b2cbfa 100644
--- a/include/linux/kasan.h
+++ b/include/linux/kasan.h
@@ -302,7 +302,7 @@ static inline void kasan_kfree_large(void *ptr, unsigned long ip) {}

#endif /* CONFIG_KASAN */

-#if defined(CONFIG_KASAN) && CONFIG_KASAN_STACK
+#if defined(CONFIG_KASAN) && defined(CONFIG_KASAN_STACK)
void kasan_unpoison_task_stack(struct task_struct *task);
#else
static inline void kasan_unpoison_task_stack(struct task_struct *task) {}
diff --git a/lib/Kconfig.kasan b/lib/Kconfig.kasan
index f5fa4ba126bf..59de74293454 100644
--- a/lib/Kconfig.kasan
+++ b/lib/Kconfig.kasan
@@ -138,9 +138,11 @@ config KASAN_INLINE

endchoice

-config KASAN_STACK_ENABLE
- bool "Enable stack instrumentation (unsafe)" if CC_IS_CLANG && !COMPILE_TEST
+config KASAN_STACK
+ bool "Enable stack instrumentation (unsafe)"
depends on KASAN_GENERIC || KASAN_SW_TAGS
+ default y if CC_IS_GCC
+ default n if CC_IS_CLANG
help
The LLVM stack address sanitizer has a know problem that
causes excessive stack usage in a lot of functions, see
@@ -154,11 +156,6 @@ config KASAN_STACK_ENABLE
CONFIG_COMPILE_TEST. On gcc it is assumed to always be safe
to use and enabled by default.

-config KASAN_STACK
- int
- default 1 if KASAN_STACK_ENABLE || CC_IS_GCC
- default 0
-
config KASAN_SW_TAGS_IDENTIFY
bool "Enable memory corruption identification"
depends on KASAN_SW_TAGS
diff --git a/mm/kasan/common.c b/mm/kasan/common.c
index 38ba2aecd8f4..02ec7f81dc16 100644
--- a/mm/kasan/common.c
+++ b/mm/kasan/common.c
@@ -63,7 +63,7 @@ void __kasan_unpoison_range(const void *address, size_t size)
unpoison_range(address, size);
}

-#if CONFIG_KASAN_STACK
+#if defined(CONFIG_KASAN_STACK)
/* Unpoison the entire stack for a task. */
void kasan_unpoison_task_stack(struct task_struct *task)
{
diff --git a/mm/kasan/kasan.h b/mm/kasan/kasan.h
index cc4d9e1d49b1..bdfdb1cff653 100644
--- a/mm/kasan/kasan.h
+++ b/mm/kasan/kasan.h
@@ -224,7 +224,7 @@ void *find_first_bad_addr(void *addr, size_t size);
const char *get_bug_type(struct kasan_access_info *info);
void metadata_fetch_row(char *buffer, void *row);

-#if defined(CONFIG_KASAN_GENERIC) && CONFIG_KASAN_STACK
+#if defined(CONFIG_KASAN_GENERIC) && defined(CONFIG_KASAN_STACK)
void print_address_stack_frame(const void *addr);
#else
static inline void print_address_stack_frame(const void *addr) { }
diff --git a/mm/kasan/report_generic.c b/mm/kasan/report_generic.c
index 8a9c889872da..137a1dba1978 100644
--- a/mm/kasan/report_generic.c
+++ b/mm/kasan/report_generic.c
@@ -128,7 +128,7 @@ void metadata_fetch_row(char *buffer, void *row)
memcpy(buffer, kasan_mem_to_shadow(row), META_BYTES_PER_ROW);
}

-#if CONFIG_KASAN_STACK
+#if defined(CONFIG_KASAN_STACK)
static bool __must_check tokenize_frame_descr(const char **frame_descr,
char *token, size_t max_tok_len,
unsigned long *value)
diff --git a/scripts/Makefile.kasan b/scripts/Makefile.kasan
index 1e000cc2e7b4..abf231d209b1 100644
--- a/scripts/Makefile.kasan
+++ b/scripts/Makefile.kasan
@@ -2,6 +2,12 @@
CFLAGS_KASAN_NOSANITIZE := -fno-builtin
KASAN_SHADOW_OFFSET ?= $(CONFIG_KASAN_SHADOW_OFFSET)

+ifdef CONFIG_KASAN_STACK
+ stack_enable := 1
+else
+ stack_enable := 0
+endif
+
ifdef CONFIG_KASAN_GENERIC

ifdef CONFIG_KASAN_INLINE
@@ -27,7 +33,7 @@ else
CFLAGS_KASAN := $(CFLAGS_KASAN_SHADOW) \
$(call cc-param,asan-globals=1) \
$(call cc-param,asan-instrumentation-with-call-threshold=$(call_threshold)) \
- $(call cc-param,asan-stack=$(CONFIG_KASAN_STACK)) \
+ $(call cc-param,asan-stack=$(stack_enable)) \
$(call cc-param,asan-instrument-allocas=1)
endif

@@ -42,7 +48,7 @@ else
endif

CFLAGS_KASAN := -fsanitize=kernel-hwaddress \
- -mllvm -hwasan-instrument-stack=$(CONFIG_KASAN_STACK) \
+ -mllvm -hwasan-instrument-stack=$(stack_enable) \
-mllvm -hwasan-use-short-granules=0 \
$(instrumentation_flags)

--
2.18.0

[Walter Wu]

CONFIG_KASAN_STACK and CONFIG_KASAN_STACK_ENABLE both enable KASAN stack

instrumentation, but we should only need one config, so that we remove
CONFIG_KASAN_STACK_ENABLE and make CONFIG_KASAN_STACK workable. see [1].

When enable KASAN stack instrumentation, then for gcc we could do no

prompt and default value y, and for clang prompt and default value n.

[1]: https://bugzilla.kernel.org/show_bug.cgi?id=210221

Signed-off-by: Walter Wu <walter...@mediatek.com>
Suggested-by: Dmitry Vyukov <dvy...@google.com>
Cc: Andrey Ryabinin <arya...@virtuozzo.com>
Cc: Dmitry Vyukov <dvy...@google.com>
Cc: Andrey Konovalov <andre...@google.com>
Cc: Alexander Potapenko <gli...@google.com>
Cc: Andrew Morton <ak...@linux-foundation.org>
---

v2: make commit log to be more readable.

[Nathan Chancellor]

You are effectively undoing commits 6baec880d7a5 ("kasan: turn off
asan-stack for clang-8 and earlier") and ebb6d35a74ce ("kasan: remove
clang version check for KASAN_STACK") with this change. This change
should still remain around so that all{mod,yes}config remain mostly
clean for clang builds. This should not change anything from the user's
perspective because this option was never user selectable for GCC and
the default y keeps it on.

> +config KASAN_STACK
> + bool "Enable stack instrumentation (unsafe)"
> depends on KASAN_GENERIC || KASAN_SW_TAGS
> + default y if CC_IS_GCC
> + default n if CC_IS_CLANG

This is implied and can be removed.

> help
> The LLVM stack address sanitizer has a know problem that
> causes excessive stack usage in a lot of functions, see
> @@ -154,11 +156,6 @@ config KASAN_STACK_ENABLE
> CONFIG_COMPILE_TEST. On gcc it is assumed to always be safe
> to use and enabled by default.
>
> -config KASAN_STACK
> - int
> - default 1 if KASAN_STACK_ENABLE || CC_IS_GCC
> - default 0
> -
> config KASAN_SW_TAGS_IDENTIFY
> bool "Enable memory corruption identification"
> depends on KASAN_SW_TAGS
> diff --git a/mm/kasan/common.c b/mm/kasan/common.c
> index 38ba2aecd8f4..02ec7f81dc16 100644
> --- a/mm/kasan/common.c
> +++ b/mm/kasan/common.c
> @@ -63,7 +63,7 @@ void __kasan_unpoison_range(const void *address, size_t size)
> unpoison_range(address, size);
> }
>
> -#if CONFIG_KASAN_STACK
> +#if defined(CONFIG_KASAN_STACK)

Isn't '#ifdef CONFIG_...' preferred for CONFIG symbols?

[Walter Wu]

Ok, I will remain this.

Yes, I will fix in the next version. Thanks for your review.

[Walter Wu]

CONFIG_KASAN_STACK and CONFIG_KASAN_STACK_ENABLE both enable KASAN stack

instrumentation, but we should only need one config, so that we remove
CONFIG_KASAN_STACK_ENABLE and make CONFIG_KASAN_STACK workable. see [1].

When enable KASAN stack instrumentation, then for gcc we could do

no prompt and default value y, and for clang prompt and default
value n.

[1]: https://bugzilla.kernel.org/show_bug.cgi?id=210221

Signed-off-by: Walter Wu <walter...@mediatek.com>
Suggested-by: Dmitry Vyukov <dvy...@google.com>
Cc: Andrey Ryabinin <arya...@virtuozzo.com>
Cc: Dmitry Vyukov <dvy...@google.com>
Cc: Andrey Konovalov <andre...@google.com>
Cc: Alexander Potapenko <gli...@google.com>
Cc: Andrew Morton <ak...@linux-foundation.org>

Cc: Nathan Chancellor <natecha...@gmail.com>

---

v2: make commit log to be more readable.

v3: remain CONFIG_KASAN_STACK_ENABLE setting
fix the pre-processors syntax

---
arch/arm64/kernel/sleep.S | 2 +-
arch/x86/kernel/acpi/wakeup_64.S | 2 +-
include/linux/kasan.h | 2 +-

lib/Kconfig.kasan | 8 ++------

mm/kasan/common.c | 2 +-
mm/kasan/kasan.h | 2 +-
mm/kasan/report_generic.c | 2 +-
scripts/Makefile.kasan | 10 ++++++++--

8 files changed, 16 insertions(+), 14 deletions(-)

index f5fa4ba126bf..fde82ec85f8f 100644
--- a/lib/Kconfig.kasan
+++ b/lib/Kconfig.kasan
@@ -138,9 +138,10 @@ config KASAN_INLINE

endchoice

-config KASAN_STACK_ENABLE
+config KASAN_STACK

bool "Enable stack instrumentation (unsafe)" if CC_IS_CLANG && !COMPILE_TEST

depends on KASAN_GENERIC || KASAN_SW_TAGS
+ default y if CC_IS_GCC

help
The LLVM stack address sanitizer has a know problem that
causes excessive stack usage in a lot of functions, see

@@ -154,11 +155,6 @@ config KASAN_STACK_ENABLE

CONFIG_COMPILE_TEST. On gcc it is assumed to always be safe
to use and enabled by default.

-config KASAN_STACK
- int
- default 1 if KASAN_STACK_ENABLE || CC_IS_GCC
- default 0
-
config KASAN_SW_TAGS_IDENTIFY
bool "Enable memory corruption identification"
depends on KASAN_SW_TAGS
diff --git a/mm/kasan/common.c b/mm/kasan/common.c

index 38ba2aecd8f4..bf8b073eed62 100644

--- a/mm/kasan/common.c
+++ b/mm/kasan/common.c
@@ -63,7 +63,7 @@ void __kasan_unpoison_range(const void *address, size_t size)
unpoison_range(address, size);
}

-#if CONFIG_KASAN_STACK

+#ifdef CONFIG_KASAN_STACK

/* Unpoison the entire stack for a task. */
void kasan_unpoison_task_stack(struct task_struct *task)
{
diff --git a/mm/kasan/kasan.h b/mm/kasan/kasan.h
index cc4d9e1d49b1..bdfdb1cff653 100644
--- a/mm/kasan/kasan.h
+++ b/mm/kasan/kasan.h
@@ -224,7 +224,7 @@ void *find_first_bad_addr(void *addr, size_t size);
const char *get_bug_type(struct kasan_access_info *info);
void metadata_fetch_row(char *buffer, void *row);

-#if defined(CONFIG_KASAN_GENERIC) && CONFIG_KASAN_STACK
+#if defined(CONFIG_KASAN_GENERIC) && defined(CONFIG_KASAN_STACK)
void print_address_stack_frame(const void *addr);
#else
static inline void print_address_stack_frame(const void *addr) { }
diff --git a/mm/kasan/report_generic.c b/mm/kasan/report_generic.c

index 8a9c889872da..4e16518d9877 100644

--- a/mm/kasan/report_generic.c
+++ b/mm/kasan/report_generic.c
@@ -128,7 +128,7 @@ void metadata_fetch_row(char *buffer, void *row)
memcpy(buffer, kasan_mem_to_shadow(row), META_BYTES_PER_ROW);
}

-#if CONFIG_KASAN_STACK

+#ifdef CONFIG_KASAN_STACK

[Nathan Chancellor]

On Fri, Jan 08, 2021 at 12:09:40PM +0800, Walter Wu wrote:
> CONFIG_KASAN_STACK and CONFIG_KASAN_STACK_ENABLE both enable KASAN stack
> instrumentation, but we should only need one config, so that we remove
> CONFIG_KASAN_STACK_ENABLE and make CONFIG_KASAN_STACK workable. see [1].
>
> When enable KASAN stack instrumentation, then for gcc we could do
> no prompt and default value y, and for clang prompt and default
> value n.
>
> [1]: https://bugzilla.kernel.org/show_bug.cgi?id=210221
>
> Signed-off-by: Walter Wu <walter...@mediatek.com>
> Suggested-by: Dmitry Vyukov <dvy...@google.com>
> Cc: Andrey Ryabinin <arya...@virtuozzo.com>
> Cc: Dmitry Vyukov <dvy...@google.com>
> Cc: Andrey Konovalov <andre...@google.com>
> Cc: Alexander Potapenko <gli...@google.com>
> Cc: Andrew Morton <ak...@linux-foundation.org>
> Cc: Nathan Chancellor <natecha...@gmail.com>

Reviewed-by: Nathan Chancellor <natecha...@gmail.com>

[Andrey Konovalov]

AFAIR, Arnd wanted to avoid having KASAN_STACK to be enabled by
default when compiling with Clang, since Clang instrumentation leads
to very large kernel stacks, which, in turn, lead to compile-time
warnings. What I don't remember is why there are two configs.

Arnd, is that correct? What was the reason behind having two configs?

[Arnd Bergmann]

On Fri, Jan 8, 2021 at 7:56 PM Andrey Konovalov <andre...@google.com> wrote:
> On Fri, Jan 8, 2021 at 5:09 AM Walter Wu <walter...@mediatek.com> wrote:

> > @@ -2,6 +2,12 @@
> > CFLAGS_KASAN_NOSANITIZE := -fno-builtin
> > KASAN_SHADOW_OFFSET ?= $(CONFIG_KASAN_SHADOW_OFFSET)
> >
> > +ifdef CONFIG_KASAN_STACK
> > + stack_enable := 1
> > +else
> > + stack_enable := 0
> > +endif
> > +
>

> AFAIR, Arnd wanted to avoid having KASAN_STACK to be enabled by
> default when compiling with Clang, since Clang instrumentation leads
> to very large kernel stacks, which, in turn, lead to compile-time
> warnings. What I don't remember is why there are two configs.
>
> Arnd, is that correct? What was the reason behind having two configs?

I think I just considered it cleaner than defining the extra variable in the
Makefile at the time, as this was the only place that referenced
CONFIG_KASAN_STACK.

The '#if CONFIG_KASAN_STACK' (rather than #ifdef) that got added
later do make my version more confusing though, so I agree that
Walter's second patch improves it.

Acked-by: Arnd Bergmann <ar...@arndb.de>

On a related note: do you have any hope that clang will ever fix
https://bugs.llvm.org/show_bug.cgi?id=38809 and KASAN_STACK
can be enabled by default on clang without risking stack
overflows?

Arnd

[Andrey Konovalov]

On Fri, Jan 8, 2021 at 9:31 PM Arnd Bergmann <ar...@kernel.org> wrote:
>
> On Fri, Jan 8, 2021 at 7:56 PM Andrey Konovalov <andre...@google.com> wrote:
> > On Fri, Jan 8, 2021 at 5:09 AM Walter Wu <walter...@mediatek.com> wrote:
>
> > > @@ -2,6 +2,12 @@
> > > CFLAGS_KASAN_NOSANITIZE := -fno-builtin
> > > KASAN_SHADOW_OFFSET ?= $(CONFIG_KASAN_SHADOW_OFFSET)
> > >
> > > +ifdef CONFIG_KASAN_STACK
> > > + stack_enable := 1
> > > +else
> > > + stack_enable := 0
> > > +endif
> > > +
> >
> > AFAIR, Arnd wanted to avoid having KASAN_STACK to be enabled by
> > default when compiling with Clang, since Clang instrumentation leads
> > to very large kernel stacks, which, in turn, lead to compile-time
> > warnings. What I don't remember is why there are two configs.
> >
> > Arnd, is that correct? What was the reason behind having two configs?
>
> I think I just considered it cleaner than defining the extra variable in the
> Makefile at the time, as this was the only place that referenced
> CONFIG_KASAN_STACK.
>
> The '#if CONFIG_KASAN_STACK' (rather than #ifdef) that got added
> later do make my version more confusing though, so I agree that
> Walter's second patch improves it.
>
> Acked-by: Arnd Bergmann <ar...@arndb.de>

Got it, thanks!

> On a related note: do you have any hope that clang will ever fix
> https://bugs.llvm.org/show_bug.cgi?id=38809 and KASAN_STACK
> can be enabled by default on clang without risking stack
> overflows?

Not sure :(

I've filed this on KASAN bugzilla to not forget:
https://bugzilla.kernel.org/show_bug.cgi?id=211139

[Andrey Konovalov]

On Fri, Jan 8, 2021 at 5:09 AM Walter Wu <walter...@mediatek.com> wrote:
>

Does this syntax mean that KASAN_STACK is only present for
CC_IS_CLANG? Or that it can only be disabled for CC_IS_CLANG?

Anyway, I think it's better to 1. allow to control KASAN_STACK
regardless of the compiler (as it was possible before), and 2. avoid
this "bool ... if ..." syntax as it's confusing.

> --
> 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/20210108040940.1138-1-walter-zh.wu%40mediatek.com.

[Nathan Chancellor]

It means that the option can only be disabled for clang.

> Anyway, I think it's better to 1. allow to control KASAN_STACK
> regardless of the compiler (as it was possible before), and 2. avoid

It has never been possible to control KASAN_STACK for GCC because of the
bool ... if ... syntax. This patch does not change that logic. Making it
possible to control KASAN_STACK with GCC seems fine but that is going to
be a new change that would probably be suited for a new patch on top of
this one.

[Andrey Konovalov]

On Mon, Jan 11, 2021 at 7:59 PM Nathan Chancellor
<natecha...@gmail.com> wrote:
>
> > > -config KASAN_STACK_ENABLE
> > > +config KASAN_STACK
> > > bool "Enable stack instrumentation (unsafe)" if CC_IS_CLANG && !COMPILE_TEST
> >
> > Does this syntax mean that KASAN_STACK is only present for
> > CC_IS_CLANG? Or that it can only be disabled for CC_IS_CLANG?
>
> It means that the option can only be disabled for clang.

OK, got it.

> > Anyway, I think it's better to 1. allow to control KASAN_STACK
> > regardless of the compiler (as it was possible before), and 2. avoid
>
> It has never been possible to control KASAN_STACK for GCC because of the
> bool ... if ... syntax. This patch does not change that logic. Making it
> possible to control KASAN_STACK with GCC seems fine but that is going to
> be a new change that would probably be suited for a new patch on top of
> this one.

The if syntax was never applied to KASAN_STACK, only to
KASAN_STACK_ENABLE, so it should have been possible (although I've
never specifically tried it).

[Nathan Chancellor]

CONFIG_KASAN_STACK was not a user selectable symbol so it was always 1
for GCC.

Cheers,
Nathan

[Andrey Konovalov]

On Mon, Jan 11, 2021 at 8:11 PM Nathan Chancellor

Ah, indeed.

Thanks for the clarification!

[Andrey Konovalov]

On Fri, Jan 8, 2021 at 5:09 AM Walter Wu <walter...@mediatek.com> wrote:
>

Reviewed-by: Andrey Konovalov <andre...@google.com>