[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CA+fCnZczFuOo0sxcrvihzJY_j0JQmH26J=4uMaz7-bsqnhakzg@mail.gmail.com>
Date: Fri, 8 Apr 2022 14:49:59 +0200
From: Andrey Konovalov <andreyknvl@...il.com>
To: Vincenzo Frascino <vincenzo.frascino@....com>
Cc: Linux ARM <linux-arm-kernel@...ts.infradead.org>,
LKML <linux-kernel@...r.kernel.org>,
kasan-dev <kasan-dev@...glegroups.com>,
Andrey Ryabinin <ryabinin.a.a@...il.com>,
Alexander Potapenko <glider@...gle.com>,
Dmitry Vyukov <dvyukov@...gle.com>,
Andrew Morton <akpm@...ux-foundation.org>,
Catalin Marinas <catalin.marinas@....com>,
Will Deacon <will@...nel.org>
Subject: Re: [PATCH v2] kasan: Fix hw tags enablement when KUNIT tests are disabled
On Fri, Apr 8, 2022 at 2:43 PM Vincenzo Frascino
<vincenzo.frascino@....com> wrote:
>
> Kasan enables hw tags via kasan_enable_tagging() which based on the mode
> passed via kernel command line selects the correct hw backend.
> kasan_enable_tagging() is meant to be invoked indirectly via the cpu features
> framework of the architectures that support these backends.
> Currently the invocation of this function is guarded by CONFIG_KASAN_KUNIT_TEST
> which allows the enablement of the correct backend only when KUNIT tests are
> enabled in the kernel.
>
> This inconsistency was introduced in commit:
>
> ed6d74446cbf ("kasan: test: support async (again) and asymm modes for HW_TAGS")
>
> ... and prevents to enable MTE on arm64 when KUNIT tests for kasan hw_tags are
> disabled.
>
> Fix the issue making sure that the CONFIG_KASAN_KUNIT_TEST guard does not
> prevent the correct invocation of kasan_enable_tagging().
>
> Fixes: ed6d74446cbf ("kasan: test: support async (again) and asymm modes for HW_TAGS")
> Cc: Andrey Ryabinin <ryabinin.a.a@...il.com>
> Cc: Alexander Potapenko <glider@...gle.com>
> Cc: Andrey Konovalov <andreyknvl@...il.com>
> Cc: Dmitry Vyukov <dvyukov@...gle.com>
> Cc: Andrew Morton <akpm@...ux-foundation.org>
> Cc: Catalin Marinas <catalin.marinas@....com>
> Cc: Will Deacon <will@...nel.org>
> Signed-off-by: Vincenzo Frascino <vincenzo.frascino@....com>
> ---
> mm/kasan/hw_tags.c | 5 +++--
> mm/kasan/kasan.h | 10 ++++++----
> 2 files changed, 9 insertions(+), 6 deletions(-)
>
> diff --git a/mm/kasan/hw_tags.c b/mm/kasan/hw_tags.c
> index 07a76c46daa5..9e1b6544bfa8 100644
> --- a/mm/kasan/hw_tags.c
> +++ b/mm/kasan/hw_tags.c
> @@ -336,8 +336,6 @@ void __kasan_poison_vmalloc(const void *start, unsigned long size)
>
> #endif
>
> -#if IS_ENABLED(CONFIG_KASAN_KUNIT_TEST)
> -
> void kasan_enable_tagging(void)
> {
> if (kasan_arg_mode == KASAN_ARG_MODE_ASYNC)
> @@ -347,6 +345,9 @@ void kasan_enable_tagging(void)
> else
> hw_enable_tagging_sync();
> }
> +
> +#if IS_ENABLED(CONFIG_KASAN_KUNIT_TEST)
> +
> EXPORT_SYMBOL_GPL(kasan_enable_tagging);
>
> void kasan_force_async_fault(void)
> diff --git a/mm/kasan/kasan.h b/mm/kasan/kasan.h
> index d79b83d673b1..b01b4bbe0409 100644
> --- a/mm/kasan/kasan.h
> +++ b/mm/kasan/kasan.h
> @@ -355,25 +355,27 @@ static inline const void *arch_kasan_set_tag(const void *addr, u8 tag)
> #define hw_set_mem_tag_range(addr, size, tag, init) \
> arch_set_mem_tag_range((addr), (size), (tag), (init))
>
> +void kasan_enable_tagging(void);
> +
> #else /* CONFIG_KASAN_HW_TAGS */
>
> #define hw_enable_tagging_sync()
> #define hw_enable_tagging_async()
> #define hw_enable_tagging_asymm()
>
> +static inline void kasan_enable_tagging(void) { }
> +
> #endif /* CONFIG_KASAN_HW_TAGS */
>
> #if defined(CONFIG_KASAN_HW_TAGS) && IS_ENABLED(CONFIG_KASAN_KUNIT_TEST)
>
> -void kasan_enable_tagging(void);
> void kasan_force_async_fault(void);
>
> -#else /* CONFIG_KASAN_HW_TAGS || CONFIG_KASAN_KUNIT_TEST */
> +#else /* CONFIG_KASAN_HW_TAGS && CONFIG_KASAN_KUNIT_TEST */
>
> -static inline void kasan_enable_tagging(void) { }
> static inline void kasan_force_async_fault(void) { }
>
> -#endif /* CONFIG_KASAN_HW_TAGS || CONFIG_KASAN_KUNIT_TEST */
> +#endif /* CONFIG_KASAN_HW_TAGS && CONFIG_KASAN_KUNIT_TEST */
>
> #ifdef CONFIG_KASAN_SW_TAGS
> u8 kasan_random_tag(void);
> --
> 2.35.1
>
Reviewed-by: Andrey Konovalov <andreyknvl@...il.com>
Thanks!
Powered by blists - more mailing lists