From c886760efda836b060fee03bfce852f16f04b5b8 Mon Sep 17 00:00:00 2001 From: Arthur Hanson Rasmusson Date: Fri, 2 Oct 2026 02:07:52 -0400 Subject: gsp-phys: freeze-safe release, boot-proof READY, and guest-wiring P2/P3/P5-wave1 Host-freeze fix bundle (validated 2/2 supervised cycles): - power unlock now latches D0/D3cold by default (new gsp_phys_repower, default 0); releasing the claim can no longer power-gate the board and hard-freeze the host on resume - bounded post-rebind readiness gate before dropping the locks; bounded D0 settle/verify before the unbind - READY means a live GSP-RM session: the async worker and the module-owned join only expose READY when init-done is seen and no unload was sent; otherwise the claim is released fail-open - FWSEC reads DMEM0/IMEM back before STARTCPU and fails fast when the staging is not visible - the claim harness leaves the board latched by default and takes --restore-power to opt back into the old restore Guest-wiring: - P2: generation-keyed fn65 static-config snapshot cache plus a default-off identity overlay (new gsp_phys_static_live) for name/short-name/VBIOS; the probe cycle runs once per backend generation at READY - P3: default-off read-only RM_CONTROL forwarding whitelist (new gsp_phys_ctrl_forward) with emulated-handle translation to the adopted server handles and full emulation fallback - P5 wave 1: default-off shadow execution of guest CE pushes on the ladder channel (new gsp_phys_work_forward) via a dedicated ordered work item; the ladder stage-4 submission is refactored into a shared bounded helper and the software release path stays authoritative Live: module-owned boot to init-done, guest smoke ALL OK 64/64 with P2/P3 enabled, 514 shadow copies observed on the physical CE; all cycles released cleanly with WPR2 down and no reboot. --- kernel/main.c | 24 ++++++++++++++++++++---- 1 file changed, 20 insertions(+), 4 deletions(-) (limited to 'kernel/main.c') diff --git a/kernel/main.c b/kernel/main.c index f47ef69..3784a70 100644 --- a/kernel/main.c +++ b/kernel/main.c @@ -1452,9 +1452,23 @@ static int __init nv_vgpu_init(void) if (gsp_phys_gpu_param && !gsp_phys_mdev_owned_enabled()) { ret = gsp_phys_gpu_claim(); - if (ret) + if (ret) { pr_warn("gsp-resman-vfio: physical GPU claim failed: %d\n", ret); + } else { + /* + * G0.4 (P2 prep): expose the load-time module-owned + * claim as a READY backend session. The join path only + * marks the state and schedules the static-config + * snapshot probe (a work item); failure is non-fatal, + * module load keeps the emulation path. The claim + * itself is still released at rmmod. + */ + ret = gsp_phys_backend_start(); + if (ret) + pr_warn("gsp-resman-vfio: physical backend start failed: %d\n", + ret); + } } else if (gsp_phys_gpu_param) { pr_info("gsp-resman-vfio: gsp_phys_mdev_owned=1: physical GPU claim+boot deferred to the first mdev create\n"); } else if (gsp_phys_mdev_owned_enabled()) { @@ -1518,10 +1532,12 @@ static void __exit nv_vgpu_exit(void) gsp_phys_backend_begin_stop(); } else { /* - * Module-owned claim (gsp_phys_gpu=1, mdev_owned=0): keep - * today's position relative to the debugfs teardown and the - * emulation threads; module-owned sessions never forward live. + * Module-owned claim (gsp_phys_gpu=1, mdev_owned=0): flush the + * static-config probe work item (begin_stop's only effect for a + * module-owned session), then keep today's release position + * relative to the debugfs teardown and the emulation threads. */ + gsp_phys_backend_begin_stop(); gsp_phys_gpu_release(); } -- cgit v1.2.3