[PATCH] code-provenance: add exceptions

Michael S. Tsirkin posted 1 patch 4 days, 18 hours ago
Patches applied successfully (tree, apply log)
git fetch https://github.com/patchew-project/qemu tags/patchew/0bb3f577eff40a6a834d14db94a1612a2108ec67.1779748353.git.mst@redhat.com
Maintainers: Pierrick Bouvier <pierrick.bouvier@oss.qualcomm.com>
docs/devel/code-provenance.rst | 19 +++++++++++++++++++
1 file changed, 19 insertions(+)
[PATCH] code-provenance: add exceptions
Posted by Michael S. Tsirkin 4 days, 18 hours ago
As a first step towards moving with the times, let's add
exceptions. Let's also do what linux does and ask for tags
on ai generated code.

This is not going far enough imho but will hopefully be
noncontroversial.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
 docs/devel/code-provenance.rst | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/docs/devel/code-provenance.rst b/docs/devel/code-provenance.rst
index 4e6a9afe0d..c3ed7f5e69 100644
--- a/docs/devel/code-provenance.rst
+++ b/docs/devel/code-provenance.rst
@@ -347,3 +347,22 @@ requirements for contribution.  In particular, the "Signed-off-by"
 label in a patch submission is a statement that the author takes
 responsibility for the entire contents of the patch, including any parts
 that were generated or assisted by AI tools or other tools.
+
+1. When AI generated part of the contribution is not creative in nature
+   (so not even copyrighteable).
+
+2. When the creative part of the contribution is clearly QEMU-specific
+   (so derived from the prompt).
+
+
+Attribution
+^^^^^^^^^^
+
+AI agents must not add Signed-off-by tags. Instead,
+contributions should include an Assisted-by tag, like this:
+
+  Assisted-by: AGENT_NAME:MODEL_VERSION
+
+For example:
+
+  Assisted-by: Claude:claude-opus-4-6
-- 
MST
Re: [PATCH] code-provenance: add exceptions
Posted by Alex Bennée 4 days, 6 hours ago
"Michael S. Tsirkin" <mst@redhat.com> writes:

> As a first step towards moving with the times, let's add
> exceptions. Let's also do what linux does and ask for tags
> on ai generated code.
>
> This is not going far enough imho but will hopefully be
> noncontroversial.
>
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
>  docs/devel/code-provenance.rst | 19 +++++++++++++++++++
>  1 file changed, 19 insertions(+)
>
> diff --git a/docs/devel/code-provenance.rst b/docs/devel/code-provenance.rst
> index 4e6a9afe0d..c3ed7f5e69 100644
> --- a/docs/devel/code-provenance.rst
> +++ b/docs/devel/code-provenance.rst
> @@ -347,3 +347,22 @@ requirements for contribution.  In particular, the "Signed-off-by"
>  label in a patch submission is a statement that the author takes
>  responsibility for the entire contents of the patch, including any parts
>  that were generated or assisted by AI tools or other tools.
> +
> +1. When AI generated part of the contribution is not creative in nature
> +   (so not even copyrighteable).
> +
> +2. When the creative part of the contribution is clearly QEMU-specific
> +   (so derived from the prompt).
> +

Before we start adding exceptions I think we need to be more explicitly
clear about the responsibility of the person who adds the s-o-b to the
patches.

Once we open the door to AI assisted contributions we need to be ready
for a potential avalanche of slop.

> +
> +Attribution
> +^^^^^^^^^^
> +
> +AI agents must not add Signed-off-by tags. Instead,
> +contributions should include an Assisted-by tag, like this:
> +
> +  Assisted-by: AGENT_NAME:MODEL_VERSION
> +
> +For example:
> +
> +  Assisted-by: Claude:claude-opus-4-6

-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro
Re: [PATCH] code-provenance: add exceptions
Posted by Peter Maydell 4 days, 8 hours ago
On Mon, 25 May 2026 at 23:35, Michael S. Tsirkin <mst@redhat.com> wrote:
>
> As a first step towards moving with the times, let's add
> exceptions. Let's also do what linux does and ask for tags
> on ai generated code.
>
> This is not going far enough imho but will hopefully be
> noncontroversial.

I think we should have the discussion first and change the
policy text second, please. Especially your

+2. When the creative part of the contribution is clearly QEMU-specific
+   (so derived from the prompt).

I don't currently feel I would agree with: I think it is
a lot broader than I would like to have us permit.

I suspect that discussion consensus is likely to land somewhere
around permitting the kind of thing you have in mind in your
first bullet point, but we will likely want to be more expansive
in describing it and giving examples of what we do and don't
consider to be in-scope.

thanks
-- PMM