Skip to content

x86: kernel clock_gettime precision assertion panics after boot registration completes #979

Description

@ryanbreen

At revision b346196295db73f9fe4b44c8f080e75217f27693, boot 4 of the first bash docker/qemu/run-boot-parallel.sh 5 batch reached the final 64-call registration marker, then panicked at kernel/src/clock_gettime_test.rs:77.

serial_kernel.txt records an elapsed reading of 1,992,843 ns between the two calls. The panic occurs before the boot-disk oracle report, so that oracle is absent and is correctly rejected. The full batch exited 1.

Evidence will be retained in docs/planning/green-program/block/serials/508/parallel-1-4/serial_kernel.txt.gz, with the batch transcript in docs/planning/green-program/block/serials/508/parallel-1-1/gate.log.gz, on branch sched/508-boot-thread-disk-wait.

Investigate whether this elapsed-time assertion distinguishes a clock-source defect from time spent outside the measured code. Do not weaken the assertion without evidence. This is a kernel self-test panic, distinct from the userspace exit described in 631. No root cause is claimed here.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions