diff options
| -rw-r--r-- | STATE-CURRENT.md | 25 | ||||
| -rw-r--r-- | docs/10-post-1.0/experimental-host-acceleration.md | 20 |
2 files changed, 40 insertions, 5 deletions
diff --git a/STATE-CURRENT.md b/STATE-CURRENT.md index fd737c7..622f997 100644 --- a/STATE-CURRENT.md +++ b/STATE-CURRENT.md @@ -550,11 +550,26 @@ progress.** ## Deferred goals - Host NVIDIA driver (`nvidia.ko`) over a virtual function — host OS - acceleration from the open-iov device model — is a **deferred design goal**, - lower priority than guest CUDA execution, multi-VM, 3D and performance; the - verdict (not possible on this platform as-is) and a first design sketch are - in - [`docs/03-architecture/12-host-vf-acceleration-future-goal.md`](docs/03-architecture/12-host-vf-acceleration-future-goal.md). + acceleration / host display from the open-iov device model — is a + **post-1.0 stretch goal we are still only considering**. We may work on it at + some point; we will probably take a crack at it because it is a great + feature, but **do not block on it, pursue it, or design around it**. It is + strictly lower priority than every baseline goal (guest CUDA execution, + multi-VM, 3D, Windows, isolation, performance, portability). It is + **never pursued automatically**; it requires a human decision once we are + certain there is no risk — which we are not certain of now — and it may be + scrapped on architectural grounds. Treat it until further notice as + **do not implement**. + - First design (`13-host-vf-acceleration-design.md`) was written and then + **reverted on 2026-10-02** after independent reviews falsified its + load-bearing mechanisms; all associated doc edits were reverted. + - The corrected record, candidate approaches, experimental gating policy and + revisit criteria are in the post-1.0 document + [`docs/10-post-1.0/experimental-host-acceleration.md`](docs/10-post-1.0/experimental-host-acceleration.md) + (intentionally not linked from the docs front door and not part of the + active reading order; don't index it and don't design toward it). + - Earlier verdict and options: + [`docs/03-architecture/12-host-vf-acceleration-future-goal.md`](docs/03-architecture/12-host-vf-acceleration-future-goal.md). ## Environment diff --git a/docs/10-post-1.0/experimental-host-acceleration.md b/docs/10-post-1.0/experimental-host-acceleration.md index c88bcaa..38b9774 100644 --- a/docs/10-post-1.0/experimental-host-acceleration.md +++ b/docs/10-post-1.0/experimental-host-acceleration.md @@ -13,6 +13,22 @@ > 2026-10-02 after independent adversarial, soundness, and VM-risk reviews > falsified the first design. The review record is kept privately. > +> **How to treat this goal (from the owner, 2026-10-02).** We are *considering* +> doing this; we may do this at some point. We will probably take a crack at it +> because it is a great feature — but **do not block on this, do not pursue it, +> and do not design around it.** It will be important to try at some point, but +> primary objectives and the other stretch goals come first. Only when we are +> certain there is no risk — which we are not certain of now — will we make a +> **human decision** to proceed. This is **never pursued automatically**. +> Consider it, until further notice, **deferred**. +> +> **Do-not-implement rule (binding).** Treat this document and the goal it +> describes as **"Do not implement"** until after *all* design goals are +> reached — and even then the goal may be scrapped on architectural grounds. +> No aspect of the software may be designed toward a forward-looking goal that +> would support this experimental feature: no new abstractions, hooks, seams, +> parameters, or "just in case" code for it, anywhere in the baseline system. +> > **Experimental policy (binding when the time comes):** any host-acceleration > work is experimental; it may only be built behind a compile-time flag and > enabled by an explicit module-load flag; it must not alter, replace, or @@ -387,3 +403,7 @@ Until then, this document is dormant. * **Policy:** do not implement unless the user explicitly asks; when implemented, it is experimental, compile-time gated, module-load gated, duplicate-path, and reversible; the baseline system is never compromised. + The goal is **never pursued automatically** — it requires an explicit human + decision after every baseline goal is complete and we are certain there is no + risk — and it may be scrapped on architectural grounds. Until then, no aspect + of the baseline software may be designed toward it. |