225 lines
9.2 KiB
C
225 lines
9.2 KiB
C
#include <stdlib.h>
|
|
#include <stdio.h>
|
|
#include <string.h>
|
|
#include <signal.h>
|
|
#include <fcntl.h>
|
|
#include <unistd.h>
|
|
#include <wait.h>
|
|
#include <time.h>
|
|
#include <errno.h>
|
|
#include <sys/time.h>
|
|
|
|
#include "mush.h"
|
|
#include "debug.h"
|
|
#include "store.h"
|
|
|
|
/*
|
|
* This is the "jobs" module for Mush.
|
|
* It maintains a table of jobs in various stages of execution, and it
|
|
* provides functions for manipulating jobs.
|
|
* Each job contains a pipeline, which is used to initialize the processes,
|
|
* pipelines, and redirections that make up the job.
|
|
* Each job has a job ID, which is an integer value that is used to identify
|
|
* that job when calling the various job manipulation functions.
|
|
*
|
|
* At any given time, a job will have one of the following status values:
|
|
* "new", "running", "completed", "aborted", "canceled".
|
|
* A newly created job starts out in with status "new".
|
|
* It changes to status "running" when the processes that make up the pipeline
|
|
* for that job have been created.
|
|
* A running job becomes "completed" at such time as all the processes in its
|
|
* pipeline have terminated successfully.
|
|
* A running job becomes "aborted" if the last process in its pipeline terminates
|
|
* with a signal that is not the result of the pipeline having been canceled.
|
|
* A running job becomes "canceled" if the jobs_cancel() function was called
|
|
* to cancel it and in addition the last process in the pipeline subsequently
|
|
* terminated with signal SIGKILL.
|
|
*
|
|
* In general, there will be other state information stored for each job,
|
|
* as required by the implementation of the various functions in this module.
|
|
*/
|
|
|
|
/**
|
|
* @brief Initialize the jobs module.
|
|
* @details This function is used to initialize the jobs module.
|
|
* It must be called exactly once, before any other functions of this
|
|
* module are called.
|
|
*
|
|
* @return 0 if initialization is successful, otherwise -1.
|
|
*/
|
|
int jobs_init(void) {
|
|
store_init();
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* @brief Finalize the jobs module.
|
|
* @details This function is used to finalize the jobs module.
|
|
* It must be called exactly once when job processing is to be terminated,
|
|
* before the program exits. It should cancel all jobs that have not
|
|
* yet terminated, wait for jobs that have been cancelled to terminate,
|
|
* and then expunge all jobs before returning.
|
|
*
|
|
* @return 0 if finalization is completely successful, otherwise -1.
|
|
*/
|
|
int jobs_fini(void) {
|
|
store_fini();
|
|
return 0;
|
|
}
|
|
|
|
/**
|
|
* @brief Print the current jobs table.
|
|
* @details This function is used to print the current contents of the jobs
|
|
* table to a specified output stream. The output should consist of one line
|
|
* per existing job. Each line should have the following format:
|
|
*
|
|
* <jobid>\t<pgid>\t<status>\t<pipeline>
|
|
*
|
|
* where <jobid> is the numeric job ID of the job, <status> is one of the
|
|
* following strings: "new", "running", "completed", "aborted", or "canceled",
|
|
* and <pipeline> is the job's pipeline, as printed by function show_pipeline()
|
|
* in the syntax module. The \t stand for TAB characters.
|
|
*
|
|
* @param file The output stream to which the job table is to be printed.
|
|
* @return 0 If the jobs table was successfully printed, -1 otherwise.
|
|
*/
|
|
int jobs_show(FILE *file) {
|
|
// TO BE IMPLEMENTED
|
|
abort();
|
|
}
|
|
|
|
/**
|
|
* @brief Create a new job to run a pipeline.
|
|
* @details This function creates a new job and starts it running a specified
|
|
* pipeline. The pipeline will consist of a "leader" process, which is the direct
|
|
* child of the process that calls this function, plus one child of the leader
|
|
* process to run each command in the pipeline. All processes in the pipeline
|
|
* should have a process group ID that is equal to the process ID of the leader.
|
|
* The leader process should wait for all of its children to terminate before
|
|
* terminating itself. The leader should return the exit status of the process
|
|
* running the last command in the pipeline as its own exit status, if that
|
|
* process terminated normally. If the last process terminated with a signal,
|
|
* then the leader should terminate via SIGABRT.
|
|
*
|
|
* If the "capture_output" flag is set for the pipeline, then the standard output
|
|
* of the last process in the pipeline should be redirected to be the same as
|
|
* the standard output of the pipeline leader, and this output should go via a
|
|
* pipe to the main Mush process, where it should be read and saved in the data
|
|
* store as the value of a variable, as described in the assignment handout.
|
|
* If "capture_output" is not set for the pipeline, but "output_file" is non-NULL,
|
|
* then the standard output of the last process in the pipeline should be redirected
|
|
* to the specified output file. If "input_file" is set for the pipeline, then
|
|
* the standard input of the process running the first command in the pipeline should
|
|
* be redirected from the specified input file.
|
|
*
|
|
* @param pline The pipeline to be run. The jobs module expects this object
|
|
* to be valid for as long as it requires, and it expects to be able to free this
|
|
* object when it is finished with it. This means that the caller should not pass
|
|
* a pipeline object that is shared with any other data structure, but rather should
|
|
* make a copy to be passed to this function.
|
|
*
|
|
* @return -1 if the pipeline could not be initialized properly, otherwise the
|
|
* value returned is the job ID assigned to the pipeline.
|
|
*/
|
|
int jobs_run(PIPELINE *pline) {
|
|
// TO BE IMPLEMENTED
|
|
abort();
|
|
}
|
|
|
|
/**
|
|
* @brief Wait for a job to terminate.
|
|
* @details This function is used to wait for the job with a specified job ID
|
|
* to terminate. A job has terminated when it has entered the COMPLETED, ABORTED,
|
|
* or CANCELED state.
|
|
*
|
|
* @param jobid The job ID of the job to wait for.
|
|
* @return the exit status of the job leader, as returned by waitpid(),
|
|
* or -1 if any error occurs that makes it impossible to wait for the specified job.
|
|
*/
|
|
int jobs_wait(int jobid) {
|
|
// TO BE IMPLEMENTED
|
|
abort();
|
|
}
|
|
|
|
/**
|
|
* @brief Poll to find out if a job has terminated.
|
|
* @details This function is used to poll whether the job with the specified ID
|
|
* has terminated. This is similar to jobs_wait(), except that this function returns
|
|
* immediately without waiting if the job has not yet terminated.
|
|
*
|
|
* @param jobid The job ID of the job to wait for.
|
|
* @return the exit status of the job leader, as returned by waitpid(), if the job
|
|
* has terminated, or -1 if the job has not yet terminated or if any other error occurs.
|
|
*/
|
|
int jobs_poll(int jobid) {
|
|
// TO BE IMPLEMENTED
|
|
abort();
|
|
}
|
|
|
|
/**
|
|
* @brief Expunge a terminated job from the jobs table.
|
|
* @details This function is used to expunge (remove) a job that has terminated from
|
|
* the jobs table, so that space in the table can be used to start some new job.
|
|
* In order to be expunged, a job must have terminated; if an attempt is made to expunge
|
|
* a job that has not yet terminated, it is an error. Any resources (exit status,
|
|
* open pipes, captured output, etc.) that were being used by the job are finalized
|
|
* and/or freed and will no longer be available.
|
|
*
|
|
* @param jobid The job ID of the job to expunge.
|
|
* @return 0 if the job was successfully expunged, -1 if the job could not be expunged.
|
|
*/
|
|
int jobs_expunge(int jobid) {
|
|
// TO BE IMPLEMENTED
|
|
abort();
|
|
}
|
|
|
|
/**
|
|
* @brief Attempt to cancel a job.
|
|
* @details This function is used to attempt to cancel a running job.
|
|
* In order to be canceled, the job must not yet have terminated and there
|
|
* must not have been any previous attempt to cancel the job.
|
|
* Cancellation is attempted by sending SIGKILL to the process group associated
|
|
* with the job. Cancellation becomes successful, and the job actually enters the canceled
|
|
* state, at such subsequent time as the job leader terminates as a result of SIGKILL.
|
|
* If after attempting cancellation, the job leader terminates other than as a result
|
|
* of SIGKILL, then cancellation is not successful and the state of the job is either
|
|
* COMPLETED or ABORTED, depending on how the job leader terminated.
|
|
*
|
|
* @param jobid The job ID of the job to cancel.
|
|
* @return 0 if cancellation was successfully initiated, -1 if the job was already
|
|
* terminated, a previous attempt had been made to cancel the job, or any other
|
|
* error occurred.
|
|
*/
|
|
int jobs_cancel(int jobid) {
|
|
// TO BE IMPLEMENTED
|
|
abort();
|
|
}
|
|
|
|
/**
|
|
* @brief Get the captured output of a job.
|
|
* @details This function is used to retrieve output that was captured from a job
|
|
* that has terminated, but that has not yet been expunged. Output is captured for a job
|
|
* when the "capture_output" flag is set for its pipeline.
|
|
*
|
|
* @param jobid The job ID of the job for which captured output is to be retrieved.
|
|
* @return The captured output, if the job has terminated and there is captured
|
|
* output available, otherwise NULL.
|
|
*/
|
|
char *jobs_get_output(int jobid) {
|
|
// TO BE IMPLEMENTED
|
|
abort();
|
|
}
|
|
|
|
/**
|
|
* @brief Pause waiting for a signal indicating a potential job status change.
|
|
* @details When this function is called it blocks until some signal has been
|
|
* received, at which point the function returns. It is used to wait for a
|
|
* potential job status change without consuming excessive amounts of CPU time.
|
|
*
|
|
* @return -1 if any error occurred, 0 otherwise.
|
|
*/
|
|
int jobs_pause(void) {
|
|
// TO BE IMPLEMENTED
|
|
abort();
|
|
}
|