[benweru]

What and why

Records the first real runs of i2c_sensor_nack and of disk_full on a Linux kernel, plus the steps to reproduce them on a Windows laptop with WSL2. Adds docs/linux-fault-hosts.md and updates the Sprint 3 brief (docs/tasks/sprint3-linux-fault-injection.md), which still said i2c_sensor_nack could not be triggered anywhere here. Docs only: no code changes, and i2c_sensor_nack stays opt-in.

How it was verified (real output)

Host: Linux Benweru 6.6.87.2-microsoft-standard-WSL2 #1 SMP PREEMPT_DYNAMIC Thu Jun 5 18:30:46 UTC 2025 x86_64, repo at a9b93f6. The i2c-stub and i2c-dev modules were built from linux-msft-wsl-6.6.87.2 with CONFIG_I2C_STUB=m. Their vermagic matches uname -r, and no kernel image was replaced.

pytest -m linux_i2c, as root with ORBITLOG_ENABLE_I2C_STUB=1:

tests/test_linux_faults.py::test_i2c_live PASSED
1 passed, 85 deselected in 0.49s

i2c_sensor_nack, seeds 0 to 7:

seed 0: triggered | OSError: [Errno 19] No such device | dmesg lines=0
seed 1: triggered | OSError: [Errno 19] No such device | dmesg lines=1
seed 2: triggered | OSError: [Errno 19] No such device | dmesg lines=0
seed 3: triggered | OSError: [Errno 19] No such device | dmesg lines=0
seed 4: triggered | OSError: [Errno 19] No such device | dmesg lines=0
seed 5: triggered | OSError: [Errno 19] No such device | dmesg lines=0
seed 6: triggered | OSError: [Errno 19] No such device | dmesg lines=1
seed 7: triggered | OSError: [Errno 19] No such device | dmesg lines=0

make dataset N=30 SEED=0 in WSL2 with cFS running:

cases: 30
by fault_class: cfs_app_crash_restart=5, corrupt_input_file=5, disk_full=4, oom_alloc_failure=4, payload_python_exception=4, sb_pipe_full=4, table_load_validation=4
by trigger_status: triggered=30
by capture: captured=30

The disk_full cases contain OSError: [Errno 28] No space left on device. Run from Windows, all 5 disk_full cases failed.

make test on Windows with this branch: 67 passed, 4 skipped, 15 deselected in 3.54s. In WSL2 (main, before this branch): 70 passed, 1 skipped, 15 deselected in 4.03s.

Known issues, stated in the docs

• i2c-stub returns Errno 19, not the article's Errno 121, and prints no no acknowledge from address kernel line. This class has the shape of the fault but not the tegra signature.
• The dmesg.txt window is unreliable: 0 or 1 lines, and the counts differ between runs. Needs a follow-up fix.
• The dataset in WSL2 used a docker shim that calls the Windows docker.exe. That setup is documented but not part of the repo.