unom/punktfunk
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases 4 Wiki Activity

feat(windows-host): activate render-adapter pin; gamepad SHM from proto (audit §4.2h/§6.1)

Browse Source 
§4.2h (C2): the host already pins the discrete GPU via IOCTL_SET_RENDER_ADAPTER on the IDD-push path; now that the pf-vdisplay driver implements it (0a7ae5e), correct the stale 'driver returns STATUS_NOT_IMPLEMENTED / STEP-4 stub' comments. Hybrid iGPU+dGPU boxes now actually pin the NVENC GPU.

§6.1 (C5): switch the host gamepad SHM consumers (inject/{dualsense,gamepad}_windows.rs) to derive size/offsets/magic/name from pf_vdisplay_proto::gamepad::{PadShm,XusbShm} via offset_of!/size_of!/helpers, instead of hand-literal OFF_*/140 — proto is now the single source of truth (driver-side switch follows with the gamepad-driver unification). The DualShock4 backend reuses the same pub(super) consts unchanged.

Verified: host clippy (nvenc) clean on the RTX box (x86_64-pc-windows-msvc).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
Enrico Bühler
2026-06-25 13:02:22 +00:00
parent 0a7ae5ef09
commit e5c9ee8327
3 changed files with 40 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/punktfunk-host/src/vdisplay/pf_vdisplay.rs
+9 -9
View File
@@ -245,16 +245,17 @@ unsafe fn create_monitor(device: isize, mode: Mode, watchdog_s: u32) -> Result<M
// SET_RENDER_ADAPTER is OPT-IN. By default we do NOT pin the render adapter — let the IDD use
// its natural adapter (Apollo-parity; avoids the cross-GPU mismatch ACCESS_LOST storm). Opt in
// with PUNKTFUNK_RENDER_ADAPTER=<name substring> or the IDD-push path (which MUST run NVENC on
// the discrete render GPU it pins here). NOTE: the pf-vdisplay driver currently returns
// STATUS_NOT_IMPLEMENTED for this IOCTL (a STEP-4 stub), so the call below is tolerated to fail.
// the discrete render GPU it pins here). The pf-vdisplay driver now IMPLEMENTS this IOCTL
// (IddCxAdapterSetRenderAdapter); a failure is still tolerated (the driver also reports its real
// render LUID in the shared header, so the host binds to the right GPU regardless).
let pinned = if std::env::var("PUNKTFUNK_RENDER_ADAPTER").is_ok() {
unsafe { resolve_render_adapter_luid() }
} else if std::env::var_os("PUNKTFUNK_IDD_PUSH").is_some() {
// P2 direct frame push: the host opens the driver's shared textures AND runs NVENC on the
// RENDER adapter, so on a hybrid box (dGPU + iGPU) it MUST be the discrete encoder GPU — an
// iGPU-rendered surface is untouchable by NVENC. pf-vdisplay HONORS SET_RENDER_ADAPTER (once
// implemented), so pin the discrete GPU; the driver also reports the resulting render LUID in
// the shared header, so the host binds correctly even if this is overridden.
// iGPU-rendered surface is untouchable by NVENC. pf-vdisplay now IMPLEMENTS
// SET_RENDER_ADAPTER, so pin the discrete GPU; the driver also reports the resulting render LUID
// in the shared header, so the host binds correctly even if this is overridden.
tracing::info!("IDD push: pinning the discrete render GPU (SET_RENDER_ADAPTER)");
unsafe { resolve_render_adapter_luid() }
} else {
@@ -270,11 +271,10 @@ unsafe fn create_monitor(device: isize, mode: Mode, watchdog_s: u32) -> Result<M
luid = format!("{:08x}:{:08x}", luid.HighPart, luid.LowPart),
"pf-vdisplay SET_RENDER_ADAPTER: pinned IDD render GPU"
),
// The driver currently stubs this IOCTL (STATUS_NOT_IMPLEMENTED) — warn + continue, do
// NOT propagate. The natural-adapter path still works (Apollo-parity).
// Non-fatal: warn + continue (do NOT propagate). The driver reports its real render LUID
// in the shared header and the host binds to that, so the natural-adapter path still works.
Err(e) => tracing::warn!(
"pf-vdisplay SET_RENDER_ADAPTER failed (driver stub / not implemented — \
continuing): {e:#}"
"pf-vdisplay SET_RENDER_ADAPTER failed (continuing on the natural adapter): {e:#}"
),
}
}