summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorArthur Hanson Rasmusson <arthur@vgpu.io>2026-10-01 08:47:09 -0400
committerArthur Hanson Rasmusson <arthur@vgpu.io>2026-10-01 08:47:09 -0400
commitb9b948ba1570c54f5279176d56717b928c23ac0a (patch)
treeb703234fec5bd9b0da9f3a9ff6427ea90f288ece
parentc3b962b384b4572a74f00ed8a00ad63be6be3cda (diff)
downloadopen-iov-b9b948ba1570c54f5279176d56717b928c23ac0a.tar.gz
open-iov-b9b948ba1570c54f5279176d56717b928c23ac0a.zip
STATE-CURRENT: FWSEC carve milestone; fail-safe teardown limit; approved reboot for restoration
-rw-r--r--STATE-CURRENT.md9
1 files changed, 9 insertions, 0 deletions
diff --git a/STATE-CURRENT.md b/STATE-CURRENT.md
index ffa2dde..131cf6e 100644
--- a/STATE-CURRENT.md
+++ b/STATE-CURRENT.md
@@ -47,6 +47,15 @@ chain succeeds.
passes end-to-end (`ALL OK`, payload 64/64) with real releases
(`faked=0`, `miss=0`, `stale=0`), confirming the recent pump/parser and
rewind changes did not regress the baseline.
+- **Physical-backend milestone (2026-10-01)**: the module-owned GPU path
+ reached the first successful secure-verification (FWSEC) carve of the
+ protected region on the real TU104 under the verified D0 power-domain
+ harness. The teardown of a FWSEC-only carve is not host-recoverable on this
+ fused part, so the GPU is left in the documented fail-safe (protected
+ region up, distribution driver deliberately not rebound) and is being
+ restored by an approved reboot; the next stage is a single-cycle real boot
+ (secure verification -> bootloader -> RISC-V start -> resource-manager
+ initialization) so the normal boot lifecycle makes teardown valid.
- **The full CUDA driver-API smoke test passes with correct data**: the client
completes `cuCtxCreate -> cuMemAlloc -> cuMemcpyHtoD -> cuMemcpyDtoH ->
cuMemFree -> cuCtxDestroy` and prints `ALL OK`, with the 64-byte payload