mirror of
https://github.com/NVIDIA/OpenShell.git
synced 2026-10-14 04:59:28 +08:00
pull-request/3521
12
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 DiveshandDrew Newberry
|
277f922d6d |
fix(sandbox): remove privileged control socket filters (#4279)
* fix(sandbox): reject local control peers when the kernel denies the ingress filter
Binding the boundary control listener attaches a classic BPF filter that drops
loopback-interface ingress. The fix for CVE-2026-53236 gates SO_ATTACH_FILTER on
CAP_NET_ADMIN, and a sandbox pod holds no capabilities at all, so the attach
returns EPERM and every sandbox dies at startup with:
bind boundary control listener: Operation not permitted (os error 1)
Granting the capability is not an option: the native Linux audit evidence
requires an empty capability set, so a sandbox that could attach the filter
would fail its own qualification instead.
Treat that EPERM as "filter unavailable" and move the check to accept time. Any
peer the workload can reach the listener from arrives over `lo`, so one end is
loopback or the two ends share an address; a supervisor runs in a different pod
and presents a different address. Every non-loopback listener applies the check
regardless of whether the filter attached, so a sandbox behaves the same on any
kernel and the filter is an extra standing layer where it is available.
The peer address comes from accept() rather than getpeername(), which fails with
ENOTCONN when a peer resets while still queued; the serve loop also treats
NotConnected, ConnectionReset and ConnectionAborted as per-connection errors so a
reset cannot stop the listener. An unavailable filter is reported as an OCSF
ConfigStateChange so operators can see which sandboxes run in this mode.
The isolation-interface filter tests now skip when the kernel denies the attach,
since they fail on every patched kernel without CAP_NET_ADMIN.
Signed-off-by: divesh <dgude@nvidia.com>
* fix(sandbox): remove privileged control socket filters
Use the control listener's peer checks on every kernel instead of attaching a packet filter. Retain workload socket confinement, the broker's control-port reservation, and reset-safe acceptance. Cover local and supervisor address classification and require accept-time rejection in the listener test.
Signed-off-by: Drew Newberry <anewberry@nvidia.com>
* docs: remove support matrix changes from filter fix
Signed-off-by: Drew Newberry <anewberry@nvidia.com>
---------
Signed-off-by: divesh <dgude@nvidia.com>
Signed-off-by: Drew Newberry <anewberry@nvidia.com>
Co-authored-by: Drew Newberry <anewberry@nvidia.com>
|
||
|
Drew Newberry
|
12cec59bf4 |
fix(sandbox): accept local connections natively on loopback-confined sockets (#4150)
* fix(sandbox): accept local connections natively on loopback-confined sockets On kernels without SECCOMP_FILTER_FLAG_WAIT_KILLABLE_RECV (RHEL 9 / RHCOS 5.14), the broker cannot safely write an accepted peer address into workload memory, so accept/accept4 with a peer-address buffer failed with EOPNOTSUPP. Static binaries and Go servers, which issue the raw syscall, could not accept connections at all. Move local acceptance into the kernel and replace per-accept inspection with standing kernel confinement: - Bind every broker-created TCP/UDP socket to the loopback device before injection and verify the binding. Accepted sockets inherit it, so they can neither receive routed ingress nor emit routed egress. - Stop notifying accept/accept4 and remove the accept workers, the SIGUSR2 accept-interrupt monitor, and the 64-accept ceiling. - Continue getpeername natively for every descriptor except relayed connections. - Deny interface-selection socket options and MSG_FASTOPEN sends from scalar syscall arguments, so confinement does not depend on the capability state of the user namespace that owns the network namespace and covers unregistered descriptors. - Drop loopback-interface ingress on the non-loopback TCP control listener before it listens, so a workload cannot reach it by reconnecting a natively accepted socket. - Mark descriptors above stdio close-on-exec in every workload pre_exec path. - Require an active confinement probe at qualification and carry it as required authenticated audit evidence. Document OpenShift 4.19 as the minimum release: RHCOS kernels for 4.16 through 4.18 are built without Landlock. Closes #4058 Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): close review gaps in native local accept confinement - Reject a TCP boundary control listener on a loopback address, including IPv4-mapped loopback. Production drivers bind the unspecified address; a loopback listener would be reachable from workload sockets the broker does not track. - Extend the confinement probe to IPv6 and prove that an accepted socket keeps its loopback binding after an AF_UNSPEC disconnect. - Pin the accepted-peer contract with tests: a loopback-bound listener admits only clients in the sandbox network namespace, and workload sockets cannot bind a non-loopback source address. - Correct the support matrix: accepted connections are bounded by per-process descriptor limits, the runtime PID limit, and the sandbox memory limit, not by a broker limit. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(sandbox): replace unsafe socket calls with socket2 and rustix Socket confinement now uses socket2 for device binding and the ingress filter, and rustix for interface lookup and AF_UNSPEC disconnect, so the module contains no unsafe code. The control-listener filter is no longer locked: no safe API exposes SO_LOCK_FILTER, and the listener descriptor never leaves the trusted sandbox process, which marks every descriptor above stdio close-on-exec before running workload code. Tests added for native accept use rustix and socket2 instead of raw libc calls. rustix issues accept4 and getpeername as raw syscalls, so the direct-syscall test keeps its meaning. The close-on-exec sweep test runs in a re-executed test process instead of a forked child. The close_range(CLOSE_RANGE_CLOEXEC) syscall in the pre_exec hook remains the only unsafe added by this branch; neither rustix nor nix wraps it. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): allowlist workload socket families The workload seccomp filter denied only AF_PACKET, AF_BLUETOOTH, and AF_VSOCK, and the broker continued socket() for every non-INET family. Several protocol families, including AF_RXRPC, AF_SMC, and AF_KCM, carry traffic over kernel-owned sockets that the broker never creates and that are not bound to loopback. Allow only AF_UNIX, AF_NETLINK (still limited to NETLINK_ROUTE), and the brokered AF_INET/AF_INET6 families, and restrict socketpair(2) to AF_UNIX. Both decisions use the scalar domain argument. The broker independently refuses non-INET families other than AF_UNIX and AF_NETLINK with EAFNOSUPPORT. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(sandbox): remove legacy read-only mode After native local accept, only two broker paths still wrote into workload memory: getpeername on relayed connections and the per-message lengths of a first sendmmsg to the DNS relay. Legacy mode existed only to refuse those writes on kernels without WAIT_KILLABLE_RECV, and the getpeername refusal broke CPython TLS on RHEL 9. The broker now never writes workload memory: - getpeername is no longer mediated and reports the kernel peer on every kernel; relayed connections report the loopback relay address. - A first send to the DNS relay pins the broker's socket copy to the relay and continues the syscall, so the kernel performs the send and writes any per-message results. Remove the listener mode, the task-memory write path and probe, and the task_memory_write, cancellation, and task_memory_writes_disabled audit evidence fields. WAIT_KILLABLE_RECV is still used when available and is reported for diagnostics, but no mediation decision depends on it. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(sandbox): drop WAIT_KILLABLE_RECV and make handlers restart-safe The broker no longer writes workload memory, so killable notification waits only reduced how often a signal restarts a notified syscall. RHEL 9 kernels never had them, so the handlers must tolerate restarts anyway. Install the listener without the flag on every kernel instead of special-casing newer ones. Handlers now check that the notification is still live immediately before each side effect (bind, connect of the retained socket, listen, and the first DNS send), and answer a restarted operation the broker already completed as the kernel would: a repeated TCP connect returns EISCONN, repeating the same UDP association succeeds, a repeat of a completed bind succeeds, and a failed relay reports its errno. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): finish teardown only when every workload descendant has exited Teardown waited only for registered process groups to disappear. A root is unregistered once reaped, so a descendant that ignored SIGTERM could outlive it while termination reported success and never sent SIGKILL, violating the bounded-termination requirement. The sandbox now becomes a child subreaper when it is not PID 1, so orphaned descendants stay in its tree and are reaped. Termination waits until no live descendant remains, and every scanned process is signalled through a pidfd after confirming its start time, so a reused PID is never signalled. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): keep a frozen workload from resuming itself Freezing stops every workload process with SIGSTOP. tgkill and rt_tgsigqueueinfo were not mediated, and mediated kill, rt_sigqueueinfo, and tkill passed SIGCONT through, so a workload process that was not yet stopped could resume the others while the supervisor recovered. Mediate tgkill and rt_tgsigqueueinfo like tkill: refuse targets in the sandbox thread group, report a thread outside the named group as missing, and continue otherwise. The boundary marks the broker frozen before stopping the workload and clears it after resuming; while frozen, workload requests to send SIGCONT fail with EPERM. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): send mediated DNS datagrams from the broker socket The native UDP send path re-ran the workload's own sendmsg, so per-message ancillary data such as IP_PKTINFO rode along; only the loopback destination contained a routing override. The broker now reads the datagram and sends it from its retained, relay-connected socket, which it builds without ancillary data, so a per-message override cannot redirect the packet. It writes nothing back into workload memory. A send carrying control data is refused with EOPNOTSUPP rather than silently stripped. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): keep sandbox control variables out of the workload environment The canonical process inherits the sandbox environment so the image's own ENV reaches the workload, then removed only a denylist of credential variables. Other variables in the reserved OPENSHELL_ namespace, such as the serialized user environment and the log level, still reached the workload. Remove every inherited OPENSHELL_ variable before applying the declared environment, and restore OPENSHELL_SANDBOX=1. The image's ordinary ENV and declared variables are unaffected; declared variables cannot use the reserved namespace. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): keep /proc read-only for GPU workloads GPU mode granted read-write access to all of /proc so CUDA's cuInit could write thread names to /proc/<pid>/task/<tid>/comm. Keep /proc read-only. Open mediation now serves a writable open of the caller's own thread comm file: the broker opens it and injects the descriptor, with no syscall continued. The kernel accepts a comm write only from the target's own thread group, so a substituted path or reused thread ID cannot rename another process's thread. Any other /proc write is left to Landlock, which denies it. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): keep the broker socket for connected UDP DNS sockets Sending DNS datagrams from the broker's socket required the broker to keep its copy, but the connect paths to the relay and to loopback still released it. glibc connects the resolver socket and then sends A and AAAA together with sendmmsg, which is mediated, so resolution failed. Release the copy on those connect paths only for TCP. Add a regression test for connect followed by sendmmsg that fails fast rather than hanging. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): keep slow loopback connects from stalling mediation The loopback connect path held the registry lock while polling a TCP connect for up to five seconds on the single notification dispatcher. A workload connecting to a busy local listener stalled every other mediated syscall, including opens and signals. Connect a duplicate of the retained socket without holding the lock. A nonblocking socket gets the native EINPROGRESS and the kernel completes the handshake on the shared socket; a blocking socket waits on a bounded worker thread. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): contain network broker handler panics A panic in a notification handler unwound the single broker thread while the health flag still read healthy, so every later blocked workload syscall hung until the sandbox was killed. Run each dispatch under catch_unwind: a panicking handler fails only that syscall with EIO and the broker keeps mediating. If the broker thread ever exits, mark it unhealthy so dependent operations fail closed instead of blocking. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): reject ancillary data on DNS sends instead of broker-sending Sending mediated DNS datagrams from the broker's own socket required keeping a broker handle for each DNS socket's whole life, which regressed real name resolution through reclaim and resource accounting that the mock-based unit tests did not exercise. Revert to continuing the kernel send, but refuse a send that carries ancillary control data (msg_controllen != 0) at read time, so a per-message routing override such as IP_PKTINFO cannot ride a mediated DNS send. A loopback destination contains an override that races the check. Full broker-side UDP mediation is left to a separate change with deployment e2e. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): address review findings across the hardening changes Fixes: - The blocking local-connect worker no longer toggles O_NONBLOCK on the open file it shares with the workload; it waits with a plain connect. - tgkill and rt_tgsigqueueinfo are notified only when they send SIGCONT, so ordinary thread signals such as Go preemption stay in the kernel. The frozen flag is re-read immediately before delivery. - A shell redirect to the caller's own thread comm file works again; O_CREAT and O_TRUNC are no-ops there and O_EXCL returns EEXIST. - The teardown scan is a linear walk and fails closed when /proc cannot be read. Simplifications: - Remove tests that need infrastructure outside the repo or prove nothing: the topology harness tests, the static-server benchmark, the direct-syscall accept test, the comm rename test without Landlock, a serde-default test, and the test-only panic hook in setsockopt. - Drop the unused Failed connect outcome, the redundant read-back after binding to loopback, redundant OPENSHELL_SANDBOX settings, and stale or duplicated comments; reuse the reserved environment prefix constant. - Tighten the support matrix and OpenShift wording. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(cli): wait for input when piped exec stdin is nonblocking Processes that inherit the same stdin share one open file description, so any of them can make it nonblocking for all. `sandbox exec` then read no input yet, got EAGAIN, and failed with "Resource temporarily unavailable (os error 11)". Parallel e2e tests inherit the runner's stdin, which made credential_gating fail intermittently. The stdin reader now blocks in poll() until input or end of file arrives instead of treating EAGAIN as fatal, and it leaves the shared descriptor's flags alone. The e2e exec helper also stops inheriting the runner's stdin, since its commands take no input. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): fix macOS lint and older-glibc test linking Import HashSet only in the Linux-only process scan, and call gettid through the raw syscall in a test, since the libc wrapper needs glibc 2.30 or newer. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs(sandbox): clarify socket peer address reporting Signed-off-by: Drew Newberry <anewberry@nvidia.com> --------- Signed-off-by: Drew Newberry <anewberry@nvidia.com> |
||
|
Drew Newberry
|
252882f37f |
feat(providers): serve sandbox config files on demand (#3832)
* feat(providers): serve sandbox config files on demand Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs(providers): defer managed file documentation Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs(providers): mark managed file api experimental Signed-off-by: Drew Newberry <anewberry@nvidia.com> * style(go): format provider profile fields Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(providers): preserve legacy environment with managed files Signed-off-by: Drew Newberry <anewberry@nvidia.com> --------- Signed-off-by: Drew Newberry <anewberry@nvidia.com> |
||
|
Drew Newberry
|
4688061882 |
fix(sandbox): deliver complete exec output before success (#3688)
* fix(sandbox): preserve exec output through channel close Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(exec): propagate output delivery failures before exit Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs(cli): clarify exec output delivery failures Signed-off-by: Drew Newberry <anewberry@nvidia.com> --------- Signed-off-by: Drew Newberry <anewberry@nvidia.com> |
||
|
Jim Meyer
|
62df64625b |
fix(identity): assess leaf and ancestor executable identities (#3633)
* fix(security): pin supplied executable identity chains Signed-off-by: Jim Meyer <jimeyer@nvidia.com> * docs(security): document executable identity chain pinning Signed-off-by: Jim Meyer <jimeyer@nvidia.com> * fix(binary-identity): compile Linux ancestry hashing Signed-off-by: Jim Meyer <jimeyer@nvidia.com> * test(binary-identity): avoid cross-label ancestry fixture Signed-off-by: Jim Meyer <jimeyer@nvidia.com> * test(e2e): choose distinct denied TCP port Fix flaky test due to sequential port assignment on MacOS Signed-off-by: Jim Meyer <jimeyer@nvidia.com> * fix(identity): bound executable evidence cache Reject identity chains atomically when the supervisor cache reaches its hard limit, and preserve existing pins without eviction. Classify malformed or conflicting evidence separately from policy denials at the staged TCP boundary. Signed-off-by: Jim Meyer <jimeyer@nvidia.com> --------- Signed-off-by: Jim Meyer <jimeyer@nvidia.com> |
||
|
Drew Newberry
|
b671171091 |
fix(sandbox): qualify task memory against workload child (#3574)
Signed-off-by: Drew Newberry <anewberry@nvidia.com> |
||
|
Akram Ben Aissi
|
293fab75d4 |
fix(sandbox): support kernels < 5.19 via seccomp WAIT_KILLABLE_RECV fallback (#3420)
* fix(sandbox): fall back to plain seccomp listener when WAIT_KILLABLE_RECV is unavailable SECCOMP_FILTER_FLAG_WAIT_KILLABLE_RECV was added in Linux 5.19. On older kernels (for example RHEL 9.x / 5.14 nodes such as RHCOS on OpenShift) the flag is rejected with EINVAL, which made the capability-free sandbox fail to start during the notification probe with "notification launcher disappeared". Install the notification listener with WAIT_KILLABLE_RECV when the kernel supports it and fall back to a plain NEW_LISTENER on EINVAL. The fallback listener records wait_killable_recv = false: its notification receive is uninterruptible, but the sandbox is otherwise fully functional. Signed-off-by: Akram <akram.benaissi@gmail.com> * fix(sandbox): address review — legacy read-only listener mode + cancellation invariant Follow-up to the PR review (GATOR-de00bfcc-01 / mrunalp): make the < 5.19 fallback cancellation-safe instead of racing broker writes. - Record an explicit ListenerMode (Killable vs LegacyReadOnly); add writes_disabled()/mode() and emit the selected mode in qualification output (seccomp_listener_mode). - Centralize task-memory output writes behind NotificationListener:: write_task_output; in LegacyReadOnly mode getpeername, accept/accept4 with a non-null address, and sendmmsg length write-backs fail closed with EOPNOTSUPP. accept with a null address, socket/connect/bind/listen/sendto/sendmsg keep working (copied inputs, scalar responses, atomic ADDFD_SEND). - Enforce the launch invariant `cancellation || task_memory_writes_disabled` in SandboxConfirmEvidence::validate() rather than dropping cancellation unconditionally; add task_memory_writes_disabled to SeccompEvidence. - Add per-path fail-closed tests (write_task_output, write_socket_addr, a real plain listener installed on a modern kernel) and confirmation-invariant tests. - Correct the flag-semantics comments and document both modes plus the reduced legacy syscall compatibility in architecture/sandbox.md. Signed-off-by: Akram <akram.benaissi@gmail.com> * docs: document legacy read-only sandbox mode on kernels before 5.19 Kernels without SECCOMP_FILTER_FLAG_WAIT_KILLABLE_RECV (< 5.19, e.g. RHEL 9.x / RHCOS 5.14) run the sandbox in a legacy read-only cancellation mode where the broker fails closed with EOPNOTSUPP on the mediated operations that write results back into workload memory (getpeername, accept/accept4 with a non-null address, sendmmsg length write-backs). Document this observable behavior and its syscall limitations in the public Fern docs: the support-matrix kernel requirements and the OpenShift runtime guidance. Signed-off-by: Akram <akram.benaissi@gmail.com> * style(sandbox): satisfy rustfmt and clippy doc_markdown Match the pinned rustfmt (Rust 1.95.0) line-wrapping for the write_task_output call, and backtick `legacy_read_only` in the qualification-report doc comment so clippy::doc_markdown (-D warnings) passes. Signed-off-by: Akram <akram.benaissi@gmail.com> * fix(sandbox): migrate task_memory_writes_disabled into backend protocol Add the `task_memory_writes_disabled` field to `SeccompEvidence` in the backend protocol contract and relax the validation from requiring `cancellation` to accepting `cancellation || task_memory_writes_disabled`. This completes the rebase migration missed by the isolation-interface refactor: the sandbox reports this field but the backend struct lacked it, and the validator rejected every pre-5.19 legacy listener. Signed-off-by: Akram <akram.benaissi@gmail.com> --------- Signed-off-by: Akram <akram.benaissi@gmail.com> |
||
|
Piotr Mlocek
|
e367d47394 |
fix(sandbox): reclaim socket descriptors before exhaustion (#3532)
* fix(sandbox): reclaim socket descriptors before exhaustion Signed-off-by: Piotr Mlocek <pmlocek@nvidia.com> * fix(sandbox): separate socket and descriptor limits Signed-off-by: Piotr Mlocek <pmlocek@nvidia.com> * fix(sandbox): account for existing broker descriptors Signed-off-by: Piotr Mlocek <pmlocek@nvidia.com> --------- Signed-off-by: Piotr Mlocek <pmlocek@nvidia.com> |
||
|
|
50230616d5 |
refactor(runtime): retire Community image dependencies (#3386)
* feat(sandbox): default to official Alpine sandbox image default_sandbox_image() now returns docker.io/library/alpine:3.22, a generic version-qualified official image, so a fresh install no longer depends on the community sandbox image catalog. All compute drivers (docker, podman, kubernetes, vm) inherit this fallback. Part of #3116. Signed-off-by: Akram Signed-off-by: Akram <akram.benaissi@gmail.com> * feat(deploy): default deployment configs to the official Alpine sandbox image Update the shared gateway default_image, Helm chart values, the standalone Kubernetes manifest, and the dev gateway task scripts to use docker.io/library/alpine:3.22 instead of the community base image, consistent with default_sandbox_image(). GPU e2e image-build base is left unchanged (CUDA needs a glibc base). Part of #3116. Signed-off-by: Akram Signed-off-by: Akram <akram.benaissi@gmail.com> * feat(driver): default to numeric non-root identity for USER-less images With the default sandbox image now Alpine, images that declare no OCI USER must start instead of being rejected. When the image declares no USER and the policy requests none, the Podman and Docker drivers now supply a numeric non-root identity (DEFAULT_SANDBOX_UID/GID = 1000) instead of rejecting, matching the numeric-identity behavior of the Kubernetes and VM drivers. The supervisor's resolved-identity path runs the sandbox as a synthesized non-root account without the account existing in the image. Images that declare a USER keep the OCI resolution path unchanged. Part of #3116. Signed-off-by: Akram <akram.benaissi@gmail.com> Signed-off-by: Evan Lezar <elezar@nvidia.com> * test(conformance): use Alpine workload image Signed-off-by: Evan Lezar <elezar@nvidia.com> * refactor(policy): drop community image /app path from default policy The restrictive default policy granted read-only access to /app, a directory that only existed in the community base image. A generic Alpine default has no /app, so remove it. Landlock best-effort already ignores absent paths; this just stops advertising a community-specific layout in the default. Part of #3116. Signed-off-by: Akram Signed-off-by: Akram <akram.benaissi@gmail.com> * docs(config): document Alpine default images Signed-off-by: Evan Lezar <elezar@nvidia.com> * fix(podman): report early sandbox termination Signed-off-by: Evan Lezar <elezar@nvidia.com> * fix(podman): initialize rootless workspace ownership Signed-off-by: Evan Lezar <elezar@nvidia.com> * fix(sandbox): qualify NVIDIA Ubuntu default Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): initialize rootful default workspace Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(sftp): add native sandbox adapter Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sftp): gate runtime helper support to Linux Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sftp): support standard OpenSSH file operations Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sftp): harden rename and special file handling Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(runtime): remove community image dependencies Signed-off-by: Drew Newberry <anewberry@nvidia.com> * test(e2e): build provider readiness tool fixture Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(e2e): use a dedicated Noble fixture for Docker tests Signed-off-by: Evan Lezar <elezar@nvidia.com> --------- Signed-off-by: Akram Signed-off-by: Akram <akram.benaissi@gmail.com> Signed-off-by: Evan Lezar <elezar@nvidia.com> Signed-off-by: Drew Newberry <anewberry@nvidia.com> Co-authored-by: Evan Lezar <elezar@nvidia.com> Co-authored-by: Drew Newberry <anewberry@nvidia.com> |
||
|
Drew Newberry
|
96c08f111b |
refactor(isolation)!: make confirmation backend-neutral (#3366)
* refactor(isolation): make confirmation backend-neutral Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): validate confirmation evidence at host boundary Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(isolation): keep fence evidence driver-owned Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation)!: validate explicit fence projections Require each compute driver to map its native evidence to individual outer-fence guarantees, and reject incomplete projections before a boundary becomes ready. Exercise the assembled remote confirmation path for invalid audit, property, generation, and digest evidence. BREAKING CHANGE: BoundaryConfig and SandboxRuntimeDescriptor use outer_fence projections rather than the earlier driver_fence representation. State written by earlier builds cannot be decoded; operators must stop and recreate affected sandboxes after upgrading. Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(isolation): clarify outer fence ownership Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): update confirmation test fixtures Signed-off-by: Drew Newberry <anewberry@nvidia.com> --------- Signed-off-by: Drew Newberry <anewberry@nvidia.com> |
||
|
Shiju
|
9708ba9999 |
feat(providers): report applied sandbox provider changes (#3391)
* feat(providers): report applied sandbox provider changes Record exact provider mutation targets in shared configuration operations. Require authenticated evidence that credentials, effective policy, and the workload launch environment have been installed before reporting readiness. Add bounded CLI and Rust SDK status and wait support, preserving ordinary revision-scoped references for existing processes. Verify new-client rotation and acknowledged detach revocation without external-stable resolver changes. Signed-off-by: Shiju <shiju@nvidia.com> * fix(providers): align readiness times with protobuf contracts Represent readiness receipts, status, and operation times with Timestamp and report intervals with Duration. Reserve the scalar field tags, update all consumers and generated bindings, and preserve timestamp presence and nanosecond identity through storage and client validation. Qualify both empty-map constructors in the Linux boundary test so its module compiles while retaining the explicit default required by Clippy. Signed-off-by: Shiju <shiju@nvidia.com> * fix(cli): preserve provider mutation storage uncertainty Recognize the gateway's exact structured storage-uncertainty reason for provider attach, detach, and update. Explain that the change may already be saved and must be reconciled before retrying, without exposing server messages or metadata. Preserve uncertainty ahead of generic retry hints. Exercise saved mutations through the CLI and verify single submission, redaction, missing receipt handling, and untrusted error-detail rejection. Document the recovery guidance for users and the public CLI skill. Signed-off-by: Shiju <shiju@nvidia.com> * fix(cli): explain denied provider profile lookups Report exact and alias profile lookup denials with fixed permission and workspace guidance. Keep backend details redacted and stop before provider mutations. Cover denied create and update calls through the CLI. Verify the complete provider list independently in the cross-workspace OIDC regression, extracting its JSON object from surrounding startup diagnostics. Signed-off-by: Shiju <shiju@nvidia.com> --------- Signed-off-by: Shiju <shiju@nvidia.com> |
||
|
Drew Newberry
|
c1f2e7189f |
feat(isolation): implement the RFC 0012 sandbox architecture (#2942)
* feat(isolation): add RFC 0012 backend contract Signed-off-by: Drew Newberry <385+drew@users.noreply.github.com> * refactor(isolation): name the interface crate explicitly Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): expose trusted host gateway Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs(agents): inventory the MXC driver Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): add mediated DNS transport Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): tighten interface error and digest contracts Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs(isolation): remove unrelated driver inventory Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): define capability-free launch contract Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): seal confirmed boundary state Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): validate confirmation for external backend implementations Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): clarify mediated DNS identity Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(isolation): generalize loopback connector Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(isolation): unify typed network mediation Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): bind launches to sandbox sessions Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(mxc): initialize extended sandbox status Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): add boundary protocol and Linux primitives Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): harden signals and separate process status from transport Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): validate remote confirmation through public contract Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): validate wire state and propagate snapshot failures Signed-off-by: Drew Newberry <anewberry@nvidia.com> * test(isolation): import owned agent specification explicitly Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs(isolation): describe mediated DNS channel Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): bound mediation attach without nested retries Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(isolation): generalize loopback protocol Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): add transport-neutral session authentication Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(isolation): separate sandbox backend protocol Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): harden runtime boundary controls Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): add terminal boundary operation Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): split supervisor and sandbox runtimes Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): harden boundary isolation and lifecycle ownership Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): reject private root redirects and adopt typed errors Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): preserve accept thread ownership on musl Signed-off-by: Drew Newberry <anewberry@nvidia.com> * test(sandbox): isolate credential probes from filtered threads Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): return retained exec exit status to independent waiters Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): bound network mediation and preserve socket authorization Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): bound control admission and retire stale mediation Signed-off-by: Drew Newberry <anewberry@nvidia.com> * ci(e2e): select migrated drivers per stack layer Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(sandbox): implement loopback connector Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(isolation): authenticate the Sandbox Protocol Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(supervisor): rotate launch-scoped authentication Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(sandbox): consume dedicated backend crate Signed-off-by: Drew Newberry <anewberry@nvidia.com> * test(sandbox): align topology session fixture Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): align projected bootstrap bundle Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(auth): validate refreshed credentials before rotation Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): fail closed across supervisor disconnects Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): repair rebased sandbox CI Signed-off-by: Drew Newberry <anewberry@nvidia.com> * build(runtime): publish separate sandbox and supervisor images Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(config): configure the sandbox runtime image Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(ci): validate sandbox binary linkage Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(isolation): use backend and runtime terminology Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(sandbox): use a scratch runtime image Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(ci): refresh schema and dependency policy Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(sandbox): bind reconnects to supervisor process Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs: align runtime split operational guidance Signed-off-by: Drew Newberry <anewberry@nvidia.com> * chore(security): document Kubernetes runtime RBAC Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): enforce runtime lifecycle invariants Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(compute): identify sandbox start generations Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(server): restore sandbox launch sessions Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): support authenticated runtime replacement Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(auth): bind sandbox session successors Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(auth): retry pending sandbox successors Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(vm): run the supervisor outside the guest workload Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(vm): use sandbox backend protocol Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): repair rebase integration Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): use unified build toolchain Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(vm): use sandbox runtime terminology Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(vm): own guest network bootstrap Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): expose guest init version Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): select native supervisor artifacts Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): guard guest init Linux symbols Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): scope Linux test imports Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): avoid guest interface casts Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): reconcile admitted sandbox identity Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): share resolved sandbox identity Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): surface host supervisor failures Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): include guest logs on supervisor exit Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): rotate and clean runtime generations Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): make sandbox starts generation-aware Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): rotate restored sandbox sessions Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(vm): keep shared paths in the base layer Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): bind sandbox session lineage Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(docker): isolate workloads behind the host supervisor Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(docker): rotate launch-scoped authentication Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(docker): use sandbox backend protocol Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(docker): use host networking for supervisor Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): preserve host gateway alias resolution Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(docker): use separate sandbox and supervisor images Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): restore startup validation after rebase Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(docker): name the sandbox runtime directly Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): narrow supervisor CA runtime storage Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): close companion isolation gaps Signed-off-by: Drew Newberry <anewberry@nvidia.com> * test(docker): align mediated network expectations Signed-off-by: Drew Newberry <anewberry@nvidia.com> * test(docker): exercise mediated network paths Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): attach supervisor to managed network Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): defer supervisor recovery until gateway is ready Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): make sandbox starts generation-aware Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): rotate restored sandbox sessions Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): preserve workloads during session rotation Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(docker): remove unrelated configuration RFC changes Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): bind sandbox session lineage Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(kubernetes): add proxy-pod isolation topology Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(kubernetes): use sandbox backend protocol Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): use stable sandbox service authority Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(kubernetes): split sandbox and supervisor images Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): adapt proxy pods to current runtime APIs Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(kubernetes): describe the single runtime placement Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(kubernetes): simplify sandbox orchestration Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): validate deployment prerequisites Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): update Trivy Helm profile inventory Signed-off-by: Drew Newberry <anewberry@nvidia.com> * test(kubernetes): update Trivy scan inventory count Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): reuse preloaded runtime images in e2e Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): type and clean runtime resources Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): make sandbox restarts recoverable Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): rotate restored sandbox sessions Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): preserve supervisor egress Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): bind sandbox session lineage Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(podman): adopt isolated sandbox and supervisor containers Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): stage bootstrap archives at named volume destinations Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(podman): rotate launch-scoped authentication Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(podman): use sandbox backend protocol Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(podman): use host networking for supervisor Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(podman): split sandbox and supervisor images Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): repair rebase integration Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(podman): name the sandbox runtime directly Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): provision supervisor CA runtime storage Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): address isolation review findings Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): inspect Debian supervisor provenance Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): use libpod-compatible tmpfs options Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): bind verified sandbox runtime binary Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): provide external driver data directory Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): start sandbox before joining user namespace Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): separate supervisor user namespace Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): make sandbox starts generation-aware Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): rotate restored sandbox sessions Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): bind sandbox session lineage Signed-off-by: Drew Newberry <anewberry@nvidia.com> * perf(isolation): add TCP and DNS benchmark harnesses Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(perf): align benchmark timing and supported protocols Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(perf): report TCP benchmark metrics accurately Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(perf): cancel failed worker startup Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(docker): build matching local supervisor image Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(podman): make local sandbox smoke test runnable Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): wire local sandbox runtime image Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(kubernetes): narrow sandbox service RBAC Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(ci): validate split runtime artifacts Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): harden runtime session handling Signed-off-by: Drew Newberry <anewberry@nvidia.com> * feat(supervisor): add standalone network proxy role Signed-off-by: Drew Newberry <anewberry@nvidia.com> * docs(rfc): remove implementation companion notes Signed-off-by: Drew Newberry <anewberry@nvidia.com> * refactor(vm): standardize runtime release name Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(vm): pin renamed runtime artifacts Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(auth): persist sandbox runtime identity Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(runtime): restore branch validation Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(isolation): reconcile main after rebase Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(network): close unframed HTTP 1.0 responses Signed-off-by: Drew Newberry <anewberry@nvidia.com> * chore(isolation): preserve upstream OCSF updates Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(security): close credential and TLS replay paths Signed-off-by: Drew Newberry <anewberry@nvidia.com> * fix(auth): make sandbox refresh retries idempotent Signed-off-by: Drew Newberry <anewberry@nvidia.com> --------- Signed-off-by: Drew Newberry <385+drew@users.noreply.github.com> Signed-off-by: Drew Newberry <anewberry@nvidia.com> |