Re: [PATCH] docs: explain effect of smp_read_barrier_depends() on modern architectures

[Stefan Hajnoczi <stefanha@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2394 bytes --]

On Fri, Apr 7, 2023, 09:41 Paolo Bonzini <pbonzini@redhat.com> wrote:

> The documentation for smp_read_barrier_depends() does not mention the
> architectures
> for which it is an optimization, for example ARM and PPC.  As a result, it
> is not
> clear to the reader why one would use it.  Relegate Alpha to a footnote
> together
> with other architectures where it is equivalent to smp_rmb().
>
> Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
> ---
>  docs/devel/atomics.rst | 11 +++++++----
>  1 file changed, 7 insertions(+), 4 deletions(-)
>

Modulo the typo that Peter mentioned:

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>

diff --git a/docs/devel/atomics.rst b/docs/devel/atomics.rst
> index 2157d3312956..8f6273c9283e 100644
> --- a/docs/devel/atomics.rst
> +++ b/docs/devel/atomics.rst
> @@ -201,10 +201,9 @@ They come in six kinds:
>    retrieves the address to which the second load will be directed),
>    the processor will guarantee that the first LOAD will appear to happen
>    before the second with respect to the other components of the system.
> -  However, this is not always true---for example, it was not true on
> -  Alpha processors.  Whenever this kind of access happens to shared
> -  memory (that is not protected by a lock), a read barrier is needed,
> -  and ``smp_read_barrier_depends()`` can be used instead of ``smp_rmb()``.
> +  Therefore, unlike ``smp_rmb()`` or ``qatomic_load_acquire()``,
> +  ``smp_read_barrier_depends()`` can be just a compiler barrier on
> +  weakly-ordered architectures such as ARM or PPC[#]_.
>
>    Note that the first load really has to have a _data_ dependency and not
>    a control dependency.  If the address for the second load is dependent
> @@ -212,6 +211,10 @@ They come in six kinds:
>    than actually loading the address itself, then it's a _control_
>    dependency and a full read barrier or better is required.
>
> +.. [#] The DEC Alpha is an exception, because
> ``smp_read_barrier_depends()``
> +   needs a processor barrier.  On strongly-ordered architectures such
> +   as x86 or s390, ``smp_rmb()`` and ``qatomic_load_acquire()`` can
> +   also be a compiler barriers.
>
>  Memory barriers and ``qatomic_load_acquire``/``qatomic_store_release`` are
>  mostly used when a data structure has one thread that is always a writer
> --
> 2.39.2
>
>
>

[-- Attachment #2: Type: text/html, Size: 3330 bytes --]