Date:   Mon, 20 Jan 2020 15:34:45 +0100
From:   Dmitry Vyukov <dvyukov@...gle.com>
To:     Marco Elver <elver@...gle.com>
Cc:     paulmck@...nel.org, Andrey Konovalov <andreyknvl@...gle.com>,
        Alexander Potapenko <glider@...gle.com>,
        kasan-dev <kasan-dev@...glegroups.com>,
        LKML <linux-kernel@...r.kernel.org>,
        Mark Rutland <mark.rutland@....com>,
        Will Deacon <will@...nel.org>,
        Peter Zijlstra <peterz@...radead.org>,
        Boqun Feng <boqun.feng@...il.com>,
        Arnd Bergmann <arnd@...db.de>,
        Al Viro <viro@...iv.linux.org.uk>,
        Christophe Leroy <christophe.leroy@....fr>,
        Daniel Axtens <dja@...ens.net>,
        Michael Ellerman <mpe@...erman.id.au>,
        Steven Rostedt <rostedt@...dmis.org>,
        Masami Hiramatsu <mhiramat@...nel.org>,
        Ingo Molnar <mingo@...nel.org>,
        Christian Brauner <christian.brauner@...ntu.com>,
        Daniel Borkmann <daniel@...earbox.net>, cyphar@...har.com,
        Kees Cook <keescook@...omium.org>,
        linux-arch <linux-arch@...r.kernel.org>
Subject: Re: [PATCH 1/5] include/linux: Add instrumented.h infrastructure

On Mon, Jan 20, 2020 at 3:19 PM Marco Elver <elver@...gle.com> wrote:
>
> This adds instrumented.h, which provides generic wrappers for memory
> access instrumentation that the compiler cannot emit for various
> sanitizers. Currently this unifies KASAN and KCSAN instrumentation. In
> future this will also include KMSAN instrumentation.
>
> Note that, copy_{to,from}_user require special instrumentation,
> providing hooks before and after the access, since we may need to know
> the actual bytes accessed (currently this is relevant for KCSAN, and is
> also relevant in future for KMSAN).

How will KMSAN instrumentation look like?

> Suggested-by: Arnd Bergmann <arnd@...db.de>
> Signed-off-by: Marco Elver <elver@...gle.com>
> ---
>  include/linux/instrumented.h | 153 +++++++++++++++++++++++++++++++++++
>  1 file changed, 153 insertions(+)
>  create mode 100644 include/linux/instrumented.h
>
> diff --git a/include/linux/instrumented.h b/include/linux/instrumented.h
> new file mode 100644
> index 000000000000..9f83c8520223
> --- /dev/null
> +++ b/include/linux/instrumented.h
> @@ -0,0 +1,153 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +
> +/*
> + * This header provides generic wrappers for memory access instrumentation that
> + * the compiler cannot emit for: KASAN, KCSAN.
> + */
> +#ifndef _LINUX_INSTRUMENTED_H
> +#define _LINUX_INSTRUMENTED_H
> +
> +#include <linux/compiler.h>
> +#include <linux/kasan-checks.h>
> +#include <linux/kcsan-checks.h>
> +#include <linux/types.h>
> +
> +/**
> + * instrument_read - instrument regular read access
> + *
> + * Instrument a regular read access. The instrumentation should be inserted
> + * before the actual read happens.
> + *
> + * @ptr address of access
> + * @size size of access
> + */
> +static __always_inline void instrument_read(const volatile void *v, size_t size)
> +{
> +       kasan_check_read(v, size);
> +       kcsan_check_read(v, size);
> +}
> +
> +/**
> + * instrument_write - instrument regular write access
> + *
> + * Instrument a regular write access. The instrumentation should be inserted
> + * before the actual write happens.
> + *
> + * @ptr address of access
> + * @size size of access
> + */
> +static __always_inline void instrument_write(const volatile void *v, size_t size)
> +{
> +       kasan_check_write(v, size);
> +       kcsan_check_write(v, size);
> +}
> +
> +/**
> + * instrument_atomic_read - instrument atomic read access
> + *
> + * Instrument an atomic read access. The instrumentation should be inserted
> + * before the actual read happens.
> + *
> + * @ptr address of access
> + * @size size of access
> + */
> +static __always_inline void instrument_atomic_read(const volatile void *v, size_t size)
> +{
> +       kasan_check_read(v, size);
> +       kcsan_check_atomic_read(v, size);
> +}
> +
> +/**
> + * instrument_atomic_write - instrument atomic write access
> + *
> + * Instrument an atomic write access. The instrumentation should be inserted
> + * before the actual write happens.
> + *
> + * @ptr address of access
> + * @size size of access
> + */
> +static __always_inline void instrument_atomic_write(const volatile void *v, size_t size)
> +{
> +       kasan_check_write(v, size);
> +       kcsan_check_atomic_write(v, size);
> +}
> +
> +/**
> + * instrument_copy_to_user_pre - instrument reads of copy_to_user
> + *
> + * Instrument reads from kernel memory, that are due to copy_to_user (and
> + * variants).
> + *
> + * The instrumentation must be inserted before the accesses. At this point the
> + * actual number of bytes accessed is not yet known.
> + *
> + * @dst destination address
> + * @size maximum access size
> + */
> +static __always_inline void
> +instrument_copy_to_user_pre(const volatile void *src, size_t size)
> +{
> +       /* Check before, to warn before potential memory corruption. */
> +       kasan_check_read(src, size);
> +}
> +
> +/**
> + * instrument_copy_to_user_post - instrument reads of copy_to_user
> + *
> + * Instrument reads from kernel memory, that are due to copy_to_user (and
> + * variants).
> + *
> + * The instrumentation must be inserted after the accesses. At this point the
> + * actual number of bytes accessed should be known.
> + *
> + * @dst destination address
> + * @size maximum access size
> + * @left number of bytes left that were not copied
> + */
> +static __always_inline void
> +instrument_copy_to_user_post(const volatile void *src, size_t size, size_t left)
> +{
> +       /* Check after, to avoid false positive if memory was not accessed. */
> +       kcsan_check_read(src, size - left);

Why don't we check the full range?
Kernel intending to copy something racy to user already looks like a
bug to me, even if user-space has that page unmapped. User-space can
always make the full range succeed. What am I missing?


> +}
> +
> +/**
> + * instrument_copy_from_user_pre - instrument writes of copy_from_user
> + *
> + * Instrument writes to kernel memory, that are due to copy_from_user (and
> + * variants).
> + *
> + * The instrumentation must be inserted before the accesses. At this point the
> + * actual number of bytes accessed is not yet known.
> + *
> + * @dst destination address
> + * @size maximum access size
> + */
> +static __always_inline void
> +instrument_copy_from_user_pre(const volatile void *dst, size_t size)
> +{
> +       /* Check before, to warn before potential memory corruption. */
> +       kasan_check_write(dst, size);
> +}
> +
> +/**
> + * instrument_copy_from_user_post - instrument writes of copy_from_user
> + *
> + * Instrument writes to kernel memory, that are due to copy_from_user (and
> + * variants).
> + *
> + * The instrumentation must be inserted after the accesses. At this point the
> + * actual number of bytes accessed should be known.
> + *
> + * @dst destination address
> + * @size maximum access size
> + * @left number of bytes left that were not copied
> + */
> +static __always_inline void
> +instrument_copy_from_user_post(const volatile void *dst, size_t size, size_t left)
> +{
> +       /* Check after, to avoid false positive if memory was not accessed. */
> +       kcsan_check_write(dst, size - left);
> +}
> +
> +#endif /* _LINUX_INSTRUMENTED_H */
> --
> 2.25.0.341.g760bfbb309-goog
>

Powered by blists - more mailing lists