runtime/src/enclave_rpc: Runtime crash when network is flaky

[matevz]

My network on the server was a bit flaky and I got

{"component":"rofl.rofl1qrl7nqwvaledwkw3n75jd7htklvsczje554u4p6k","level":"warn","module":"runtime","msg":"thread 'tokio-runtime-worker' panicked at /cargo/git/checkouts/oasis-core-5807357916a22ef1/2b4cf1b/runtime/src/enclave_rpc/sessions.rs:493:62:","provisioner":"tdx-qemu","runtime_id":"000000000000000000000000000000000000000000000000a6d1e3ebf60dff6c","runtime_name":"","ts":"2025-11-27T15:14:06.084422972Z"}
{"component":"rofl.rofl1qrl7nqwvaledwkw3n75jd7htklvsczje554u4p6k","level":"warn","module":"runtime","msg":"called `Option::unwrap()` on a `None` value","provisioner":"tdx-qemu","runtime_id":"000000000000000000000000000000000000000000000000a6d1e3ebf60dff6c","runtime_name":"","ts":"2025-11-27T15:14:06.087963576Z"}
{"component":"rofl.rofl1qrl7nqwvaledwkw3n75jd7htklvsczje554u4p6k","level":"warn","module":"runtime","msg":"note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace","provisioner":"tdx-qemu","runtime_id":"000000000000000000000000000000000000000000000000a6d1e3ebf60dff6c","runtime_name":"","ts":"2025-11-27T15:14:06.093664487Z"}

which corresponds to this line

oasis-core/runtime/src/enclave_rpc/sessions.rs

Line 493 in c35f14b

let session_meta = sessions.get(&session.session_id).unwrap();

Since this was triggered in ROFL, nothing important was lost, but I suppose we should avoid such panics in Sapphire.

[kostko]

This is interesting, the panic should not happen as calling the method requires holding the lock so the session should exist as it cannot be modified in parallel. We should investigate what is the root cause and fix this.