From ccdc1240d5df9777503a05b247109b6136d03313 Mon Sep 17 00:00:00 2001 From: Arthur Hanson Rasmusson Date: Fri, 23 Oct 2026 20:01:53 -0400 Subject: vgsp-sema r577: never let a foreign release VA claim a CE channel's tracking slot; event-record A/B knobs; GR-DEC dump deepened r575 starved-release wedge (live): an unreadable UVM CE entry's rescued stale page decoded a release to a foreign VA (0x12006c004, pay=60); the rescue paths adopted it as the channel's tracking slot (last_rel_va) and every later recovery release targeted it, while the client's real pool (0x121000000) stopped at 25 - the client spun in uvm_tracker_wait for 34 (exact-slot poke delivered it and the client completed green). For CE channels only a pool-shaped release target (0x121000000..0x122000000) may become last_rel_va; a foreign rescued release is refused (witnessed, bounded) and the pool step is delivered through the exact-slot fallback / deferred-release machinery instead. The fake's read-before-write keeps the delivery monotonic even when a legit push carries multiple releases. Also: gsp_rm_ev_{seq,nlist,cl} runtime knobs for the guest-RM event drop A/B (the r576 experiment: guest consumes every POST_EVENT record and never enters nv_post_event), and the GR-DEC launch dump now covers 256 data pages (the mail VADD fill carries n_data=969). New: tools/cuda-ab/evhold.cu - deterministic CUDA event holder/wait probe (context + cudaEventSynchronize + a hold window) for the event experiments. --- kernel/main.c | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'kernel/main.c') diff --git a/kernel/main.c b/kernel/main.c index d0eb2b4..8d02e5b 100644 --- a/kernel/main.c +++ b/kernel/main.c @@ -2812,6 +2812,17 @@ MODULE_PARM_DESC(gsp_phys_gr_dec_log, "GR launch-decision dump budget per boot ( module_param_named(gsp_rm_ev_log, gsp_rm_ev_log_max, uint, 0644); MODULE_PARM_DESC(gsp_rm_ev_log, "GSP event-post payload witness budget per boot (default 0: off)"); +/* + * r577 experiment knobs for the guest-RM event drop (runtime-writable so + * one boot can A/B). See gsp_rm_emulation.c. All default 0 = legacy. + */ +module_param_named(gsp_rm_ev_seq, gsp_rm_ev_seq, uint, 0644); +MODULE_PARM_DESC(gsp_rm_ev_seq, "event records carry the status-queue sequence as RPC sequence (default 0: legacy 0)"); +module_param_named(gsp_rm_ev_nlist, gsp_rm_ev_nlist, uint, 0644); +MODULE_PARM_DESC(gsp_rm_ev_nlist, "set bNotifyList on event records (default 0: specific-hEvent dispatch)"); +module_param_named(gsp_rm_ev_cl, gsp_rm_ev_cl, uint, 0644); +MODULE_PARM_DESC(gsp_rm_ev_cl, "override event record hClient (default 0: the allocating client)"); + module_param_named(gsp_sema_walk_trace, gsp_sema_walk_trace, uint, 0644); MODULE_PARM_DESC(gsp_sema_walk_trace, "Trace GMMU walks for unresolved high-VA operands (default 0: off)"); module_param_named(gsp_sema_winfill, gsp_sema_winfill, bool, 0644); -- cgit v1.2.3