Conversation
…very - Update sleep_on() to track which wait queue a task is sleeping on via task->waiting_on field - Update wake_up_all() to clear waiting_on field when removing task from wait queue - Update sys_kill() to properly remove task from its wait queue before re-enqueueing to runqueue - This prevents tasks from being on both wait queue and runqueue simultaneously, which was causing the shell to hang after child process execution
The issue was that sys_waitpid() was returning 0 when no zombie child existed, causing userspace waitpid() to busy-wait in an infinite loop. When SIGCHLD arrived, the default signal handler just ignored it, so the busy-wait continued. Changes: - Add waitpid_queue wait queue for processes blocked in waitpid() - Modify sys_waitpid() to sleep on waitpid_queue when no zombie exists (unless WNOHANG) - Modify do_exit() to wake_up_all() processes sleeping in waitpid_queue - Return -EINTR from sys_waitpid after waking from signal - Fix userspace waitpid() to: * Properly check WNOHANG with bitwise & (was using logical &&) * Retry on -EINTR (signal interrupted the sleep) * Handle error cases correctly Now the shell properly blocks in waitpid() and is woken up when the child exits.
The sleep_on() function had a path that would reuse an existing wait_queue_entry if the task was already on the queue. This caused issues because: 1. If the task was woken and the entry deallocated, stale pointers could remain 2. The check would return early without allocating a fresh entry 3. This could lead to the task not actually sleeping (early return) Solution: - Remove the entry-reuse optimization completely - Always allocate a fresh entry for each sleep_on() call - This ensures tasks actually sleep when added to the queue - Fixes the shell hanging after child process execution Also added missing clearing of waiting_on field when allocation fails.
- Add safety checks in scheduler_enqueue_task and scheduler_dequeue_task to detect and prevent double-enqueue/dequeue bugs - Implement __dump_runqueue function for debugging scheduler state - Remove verbose debug logging from enqueue/dequeue operations - Change log levels from DEBUG to NOTICE in scheduler, process, timer, and wait modules - Enhance shell status handling using environment variables instead of globals - Add comprehensive syslog debugging in shell for command execution and signal handling - Update VSCode editor settings for consistent formatting
- re-enqueue task on sleep timer expiry after removing wait queue entry\n- clear waiting_on when waking from sleep timeout\n- make RR selection robust when current task is dequeued\n- guard runtime stats update to runnable queued current task
- clear waiting_on when removing wait queue entries\n- enqueue task back to runqueue after wake-up if not already queued\n- prevent TASK_RUNNING tasks from being lost outside scheduler runqueue
Separate concerns: subsystems manage wait queues, scheduler manages state. wake_up_process() now only handles task state transitions and runqueue placement. Each subsystem removes its entries before calling wake_up_process().
Stopped tasks now just change state to TASK_STOPPED while remaining on the runqueue. Scheduler naturally skips non-running tasks. SIGCONT simply transitions state back to TASK_RUNNING. Eliminates unnecessary stopped_queue wait queue infrastructure.
This was
linked to
issues
Mar 6, 2026
SIGKILL cannot be caught or ignored, so stopped tasks (TASK_STOPPED) must be woken to TASK_RUNNING state when they receive SIGKILL. Without this, a stopped process receiving kill -9 would never be scheduled to process the fatal signal. This partially addresses issue #143 by ensuring stopped tasks can be killed. SIGCONT already wakes stopped tasks via handle_stop_signal().
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.