linux-kernel - Re: [PATCH 2/2] kcov: load acquire coverage count in user-space code

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CANpmjNPWzJZrAFT3-013GJhksK0jkB6n0HmF+h0hdoQUwGuxfA@mail.gmail.com>
Date: Mon, 28 Jul 2025 22:32:48 +0200
From: Marco Elver <elver@...gle.com>
To: Soham Bagchi <soham.bagchi@...h.edu>
Cc: dvyukov@...gle.com, andreyknvl@...il.com, akpm@...ux-foundation.org, 
	tglx@...utronix.de, glider@...gle.com, sohambagchi@...look.com, arnd@...db.de, 
	kasan-dev@...glegroups.com, linux-kernel@...r.kernel.org, corbet@....net, 
	workflows@...r.kernel.org, linux-doc@...r.kernel.org
Subject: Re: [PATCH 2/2] kcov: load acquire coverage count in user-space code

On Mon, 28 Jul 2025 at 20:43, Soham Bagchi <soham.bagchi@...h.edu> wrote:
>
> Updating the KCOV documentation to use a load-acquire
> operation for the first element of the shared memory
> buffer between kernel-space and user-space.
>
> The load-acquire pairs with the write memory barrier
> used in kcov_move_area()
>
> Signed-off-by: Soham Bagchi <soham.bagchi@...h.edu>
> ---
>  Documentation/dev-tools/kcov.rst | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
>
> diff --git a/Documentation/dev-tools/kcov.rst b/Documentation/dev-tools/kcov.rst
> index 6611434e2dd..46450fb46fe 100644
> --- a/Documentation/dev-tools/kcov.rst
> +++ b/Documentation/dev-tools/kcov.rst
> @@ -287,6 +287,11 @@ handle instance id.
>  The following program demonstrates using KCOV to collect coverage from both
>  local tasks spawned by the process and the global task that handles USB bus #1:
>
> +The user-space code for KCOV should also use an acquire to fetch the count
> +of coverage entries in the shared buffer. This acquire pairs with the
> +corresponding write memory barrier (smp_wmb()) on the kernel-side in
> +kcov_move_area().
> +

This new paragraph is misplaced.
You've added it after the "... handles USB bus #1:" part which clearly
should be right before the code (note the colon).

Why not add what you wrote here as a block-comment (similar in style
to comment above the sleep()) right above the __atomic_load_n below? I
think those details probably don't quite belong into the high level
text, but the detailed code example.

>  .. code-block:: c
>
>      /* Same includes and defines as above. */
> @@ -361,7 +366,7 @@ local tasks spawned by the process and the global task that handles USB bus #1:
>          */
>         sleep(2);
>
> -       n = __atomic_load_n(&cover[0], __ATOMIC_RELAXED);
> +       n = __atomic_load_n(&cover[0], __ATOMIC_ACQUIRE);
>         for (i = 0; i < n; i++)
>                 printf("0x%lx\n", cover[i + 1]);
>         if (ioctl(fd, KCOV_DISABLE, 0))
> --
> 2.34.1
>