diff options
Diffstat (limited to 'STATE-CURRENT.md')
| -rw-r--r-- | STATE-CURRENT.md | 41 |
1 files changed, 25 insertions, 16 deletions
diff --git a/STATE-CURRENT.md b/STATE-CURRENT.md index 6712b77..4846c24 100644 --- a/STATE-CURRENT.md +++ b/STATE-CURRENT.md @@ -1,6 +1,6 @@ # Open-IOV Current State -> **Snapshot: 2026-09-30 14:02.** This is a dated public status page, not live +> **Snapshot: 2026-09-30 15:48.** This is a dated public status page, not live > state; it is refreshed on every progress check. Live working notes are > maintained outside this repository; the public known-knowns/unknowns document > is authoritative here: @@ -30,7 +30,7 @@ meant to execute on the physical GPU. | CUDA userspace: `cuInit`, `cuDeviceGetCount`, `cuDeviceGet` | ✅ | | `cuCtxCreate` returns | ✅ (the whole smoke test now completes) | | `cuMemAlloc` + `cuMemcpy` roundtrip (`ALL OK` in the smoke test) | ✅ payload roundtrip verified 64/64 in the guest | -| Full `nvidia-smi` output | ❌ next: run real CUDA workloads in the guest | +| Full `nvidia-smi` output / kernel launch | 🟡 the static runtime now initializes (`rc=0`) and queues a kernel launch; execution awaits the backend | | CUDA kernel execution (`vectorAdd`) | ❌ not started (forwarding layer absent) | | Graphics (3D), video | ❌ | | Faults, multi-version, Windows, migration, multi-tenancy, performance | ❌ | @@ -68,19 +68,22 @@ chain succeeds. the copy-engine decoder consults pinned translations before the scanned map, so every push-data operation executes (`memset=19473`, `ringops=10551` live). -- **Runtime client status**: the static-runtime `vectorAdd` client is now past - the registration/tracker wall - the host's releases land, UVM observes the - completed wait and registration returns success. It still returns an - initialization error a few hundred milliseconds later, with no kernel-side - error; that post-registration failure is in libcuda and is being traced. -- **One real emulation bug fixed**: the BAR-info control replied with the - wrong parameter size, so the guest's reply cache rejected it; the canonical - struct is now returned. One cache-set mismatch remains once per boot and is - being identified. -- Active work: identify the remaining reply-cache mismatch, capture the - runtime client's post-registration stack if needed, and close the - device-address walk gap for the runtime's own buffers before attempting - allocation/kernel launches. +- **The runtime client now initializes and reaches kernel launch**: with a + static-runtime `vectorAdd` client, `cudaGetDeviceCount`, `cudaSetDevice`, + `cudaFree`, three `cudaMalloc`, the HtoD copies and `cudaGetLastError` all + return success. The client then parks in `cudaDeviceSynchronize` - the + launch is queued but never **executed**, because real execution requires + the physical-GPU backend (the next project). +- **What fixed it**: the failing call was never a kernel call - it was the + clock-domain probe inside libcuda (two device attributes) whose three + legacy clock controls were answered with zeros; they are now populated + with exactly the fields the helper checks. +- **Caveat**: the rc=0 init chain above was obtained in fast-release mode; + under real releases the runtime-init path still needs the pin/translation + work (the smoke-test baseline is already correct under real releases). +- Active work: (1) real-release delivery for the runtime init (pin/translation + work); (2) decode the queued compute-channel pushes into real submissions on + the physical GPU (the execution backend); then the synchronize/DtoH verify. - Observability: a size-keyed copy watch, a VRAM shadow dump/search, and an operation log are now module parameters, and the test's output survives guest-agent stalls via the serial console. @@ -137,7 +140,9 @@ Status tags: `[done]`, `[frontier]` (actively worked), `[partial]`, v [done] ------------------------------------------------> cuMemAlloc + cuMemcpy roundtrip | (passes end-to-end; payload 64/64 correct) - [not yet] ---------------------------------------------> full nvidia-smi output / kernel launch + [partial] ---------------------------------------------> full nvidia-smi output / kernel launch + | (runtime init rc=0; launch queued; + | execution awaits the backend) ``` Work execution (host service vs client synchronization): @@ -146,6 +151,8 @@ Work execution (host service vs client synchronization): [done] host decodes guest pushes (method streams): memsets, copies, releases - CE pushes execute at scale; the compute-class inline-DMA HtoD and the copy-engine-on-compute DtoH both execute; payload 64/64 + - the runtime client now initializes (rc=0) and queues a kernel + launch; the queued compute pushes decode but are not executed [partial] GPU VA -> guest PA translation: page-table walk fixed and succeeding; server-owned page-table population still in progress [not yet] real work forwarded to the physical GPU (host-RM proxy or direct submission) @@ -173,6 +180,8 @@ Recent chain of walls removed (2026-09-28 to 2026-09-30): -> cuCtxCreate OK; payload found (compute-class inline DMA) -> GMMU walk decode fixed (child/leaf encoding) -> inline HtoD lands -> DtoH copy-engine-on-compute allowed -> payload roundtrip 64/64 + -> runtime client initializes rc=0 (legacy clock-domain fix) -> kernel launch + queued -> execution backend is the next wall ``` ## Tools, code paths and fixes behind this progress |