Linux 2.6.36版本内核分析之task_struct

struct task_struct {
 volatile long state; /* -1 unrunnable, 0 runnable, >0 stopped */
 void *stack;
 atomic_t usage;
 unsigned int flags; /* per process flags, defined below */
 unsigned int ptrace;

 int lock_depth;  /* BKL lock depth */

#ifdef CONFIG_SMP
#ifdef __ARCH_WANT_UNLOCKED_CTXSW
 int oncpu;
#endif
#endif

 int prio, static_prio, normal_prio;
 unsigned int rt_priority;
 const struct sched_class *sched_class;
 struct sched_entity se;
 struct sched_rt_entity rt;

#ifdef CONFIG_PREEMPT_NOTIFIERS
 /* list of struct preempt_notifier: */
 struct hlist_head preempt_notifiers;
#endif

 /*
  * fpu_counter contains the number of consecutive context switches
  * that the FPU is used. If this is over a threshold, the lazy fpu
  * saving becomes unlazy to save the trap. This is an unsigned char
  * so that after 256 times the counter wraps and the behavior turns
  * lazy again; this to deal with bursty apps that only use FPU for
  * a short time
  */
 unsigned char fpu_counter;
#ifdef CONFIG_BLK_DEV_IO_TRACE
 unsigned int btrace_seq;
#endif

 unsigned int policy;
 cpumask_t cpus_allowed;

#ifdef CONFIG_TREE_PREEMPT_RCU
 int rcu_read_lock_nesting;
 char rcu_read_unlock_special;
 struct rcu_node *rcu_blocked_node;
 struct list_head rcu_node_entry;
#endif /* #ifdef CONFIG_TREE_PREEMPT_RCU */

#if defined(CONFIG_SCHEDSTATS) || defined(CONFIG_TASK_DELAY_ACCT)
 struct sched_info sched_info;
#endif

 struct list_head tasks;
 struct plist_node pushable_tasks;

 struct mm_struct *mm, *active_mm;
#if defined(SPLIT_RSS_COUNTING)
 struct task_rss_stat rss_stat;
#endif
/* task state */
 int exit_state;
 int exit_code, exit_signal;
 int pdeath_signal;  /*  The signal sent when the parent dies  */
 /* ??? */
 unsigned int personality;
 unsigned did_exec:1;
 unsigned in_execve:1; /* Tell the LSMs that the process is doing an
     * execve */
 unsigned in_iowait:1;

 /* Revert to default priority/policy when forking */
 unsigned sched_reset_on_fork:1;

 pid_t pid;
 pid_t tgid;

#ifdef CONFIG_CC_STACKPROTECTOR
 /* Canary value for the -fstack-protector gcc feature */
 unsigned long stack_canary;
#endif

 /*
  * pointers to (original) parent process, youngest child, younger sibling,
  * older sibling, respectively.  (p->father can be replaced with
  * p->real_parent->pid)
  */
 struct task_struct *real_parent; /* real parent process */
 struct task_struct *parent; /* recipient of SIGCHLD, wait4() reports */
 /*
  * children/sibling forms the list of my natural children
  */
 struct list_head children; /* list of my children */
 struct list_head sibling; /* linkage in my parent's children list */
 struct task_struct *group_leader; /* threadgroup leader */

 /*
  * ptraced is the list of tasks this task is using ptrace on.
  * This includes both natural children and PTRACE_ATTACH targets.
  * p->ptrace_entry is p's link on the p->parent->ptraced list.
  */
 struct list_head ptraced;
 struct list_head ptrace_entry;

 /* PID/PID hash table linkage. */
 struct pid_link pids[PIDTYPE_MAX];
 struct list_head thread_group;

 struct completion *vfork_done;  /* for vfork() */
 int __user *set_child_tid;  /* CLONE_CHILD_SETTID */
 int __user *clear_child_tid;  /* CLONE_CHILD_CLEARTID */

 cputime_t utime, stime, utimescaled, stimescaled;
 cputime_t gtime;
#ifndef CONFIG_VIRT_CPU_ACCOUNTING
 cputime_t prev_utime, prev_stime;
#endif
 unsigned long nvcsw, nivcsw; /* context switch counts */
 struct timespec start_time;   /* monotonic time */
 struct timespec real_start_time; /* boot based time */
/* mm fault and swap info: this can arguably be seen as either mm-specific or thread-specific */
 unsigned long min_flt, maj_flt;

 struct task_cputime cputime_expires;
 struct list_head cpu_timers[3];

/* process credentials */
 const struct cred *real_cred; /* objective and real subjective task
      * credentials (COW) */
 const struct cred *cred; /* effective (overridable) subjective task
      * credentials (COW) */
 struct mutex cred_guard_mutex; /* guard against foreign influences on
      * credential calculations
      * (notably. ptrace) */
 struct cred *replacement_session_keyring; /* for KEYCTL_SESSION_TO_PARENT */

 char comm[TASK_COMM_LEN]; /* executable name excluding path
         - access with [gs]et_task_comm (which lock
           it with task_lock())
         - initialized normally by setup_new_exec */
/* file system info */
 int link_count, total_link_count;
#ifdef CONFIG_SYSVIPC
/* ipc stuff */
 struct sysv_sem sysvsem;
#endif
#ifdef CONFIG_DETECT_HUNG_TASK
/* hung task detection */
 unsigned long last_switch_count;
#endif
/* CPU-specific state of this task */
 struct thread_struct thread;
/* filesystem information */
 struct fs_struct *fs;
/* open file information */
 struct files_struct *files;
/* namespaces */
 struct nsproxy *nsproxy;
/* signal handlers */
 struct signal_struct *signal;
 struct sighand_struct *sighand;

 sigset_t blocked, real_blocked;
 sigset_t saved_sigmask; /* restored if set_restore_sigmask() was used */
 struct sigpending pending;

 unsigned long sas_ss_sp;
 size_t sas_ss_size;
 int (*notifier)(void *priv);
 void *notifier_data;
 sigset_t *notifier_mask;
 struct audit_context *audit_context;
#ifdef CONFIG_AUDITSYSCALL
 uid_t loginuid;
 unsigned int sessionid;
#endif
 seccomp_t seccomp;

/* Thread group tracking */
    u32 parent_exec_id;
    u32 self_exec_id;
/* Protection of (de-)allocation: mm, files, fs, tty, keyrings, mems_allowed,
 * mempolicy */
 spinlock_t alloc_lock;

#ifdef CONFIG_GENERIC_HARDIRQS
 /* IRQ handler threads */
 struct irqaction *irqaction;
#endif

 /* Protection of the PI data structures: */
 raw_spinlock_t pi_lock;

#ifdef CONFIG_RT_MUTEXES
 /* PI waiters blocked on a rt_mutex held by this task */
 struct plist_head pi_waiters;
 /* Deadlock detection and priority inheritance handling */
 struct rt_mutex_waiter *pi_blocked_on;
#endif

#ifdef CONFIG_DEBUG_MUTEXES
 /* mutex deadlock detection */
 struct mutex_waiter *blocked_on;
#endif
#ifdef CONFIG_TRACE_IRQFLAGS
 unsigned int irq_events;
 unsigned long hardirq_enable_ip;
 unsigned long hardirq_disable_ip;
 unsigned int hardirq_enable_event;
 unsigned int hardirq_disable_event;
 int hardirqs_enabled;
 int hardirq_context;
 unsigned long softirq_disable_ip;
 unsigned long softirq_enable_ip;
 unsigned int softirq_disable_event;
 unsigned int softirq_enable_event;
 int softirqs_enabled;
 int softirq_context;
#endif
#ifdef CONFIG_LOCKDEP
# define MAX_LOCK_DEPTH 48UL
 u64 curr_chain_key;
 int lockdep_depth;
 unsigned int lockdep_recursion;
 struct held_lock held_locks[MAX_LOCK_DEPTH];
 gfp_t lockdep_reclaim_gfp;
#endif

/* journalling filesystem info */
 void *journal_info;

/* stacked block device info */
 struct bio_list *bio_list;

/* VM state */
 struct reclaim_state *reclaim_state;

 struct backing_dev_info *backing_dev_info;

 struct io_context *io_context;

 unsigned long ptrace_message;
 siginfo_t *last_siginfo; /* For ptrace use.  */
 struct task_io_accounting ioac;
#if defined(CONFIG_TASK_XACCT)
 u64 acct_rss_mem1; /* accumulated rss usage */
 u64 acct_vm_mem1; /* accumulated virtual memory usage */
 cputime_t acct_timexpd; /* stime + utime since last update */
#endif
#ifdef CONFIG_CPUSETS
 nodemask_t mems_allowed; /* Protected by alloc_lock */
 int mems_allowed_change_disable;
 int cpuset_mem_spread_rotor;
 int cpuset_slab_spread_rotor;
#endif
#ifdef CONFIG_CGROUPS
 /* Control Group info protected by css_set_lock */
 struct css_set *cgroups;
 /* cg_list protected by css_set_lock and tsk->alloc_lock */
 struct list_head cg_list;
#endif
#ifdef CONFIG_FUTEX
 struct robust_list_head __user *robust_list;
#ifdef CONFIG_COMPAT
 struct compat_robust_list_head __user *compat_robust_list;
#endif
 struct list_head pi_state_list;
 struct futex_pi_state *pi_state_cache;
#endif
#ifdef CONFIG_PERF_EVENTS
 struct perf_event_context *perf_event_ctxp;
 struct mutex perf_event_mutex;
 struct list_head perf_event_list;
#endif
#ifdef CONFIG_NUMA
 struct mempolicy *mempolicy; /* Protected by alloc_lock */
 short il_next;
#endif
 atomic_t fs_excl; /* holding fs exclusive resources */
 struct rcu_head rcu;

 /*
  * cache last used pipe for splice
  */
 struct pipe_inode_info *splice_pipe;
#ifdef CONFIG_TASK_DELAY_ACCT
 struct task_delay_info *delays;
#endif
#ifdef CONFIG_FAULT_INJECTION
 int make_it_fail;
#endif
 struct prop_local_single dirties;
#ifdef CONFIG_LATENCYTOP
 int latency_record_count;
 struct latency_record latency_record[LT_SAVECOUNT];
#endif
 /*
  * time slack values; these are used to round up poll() and
  * select() etc timeout values. These are in nanoseconds.
  */
 unsigned long timer_slack_ns;
 unsigned long default_timer_slack_ns;

 struct list_head *scm_work_list;
#ifdef CONFIG_FUNCTION_GRAPH_TRACER
 /* Index of current stored address in ret_stack */
 int curr_ret_stack;
 /* Stack of return addresses for return function tracing */
 struct ftrace_ret_stack *ret_stack;
 /* time stamp for last schedule */
 unsigned long long ftrace_timestamp;
 /*
  * Number of functions that haven't been traced
  * because of depth overrun.
  */
 atomic_t trace_overrun;
 /* Pause for the tracing */
 atomic_t tracing_graph_pause;
#endif
#ifdef CONFIG_TRACING
 /* state flags for use by tracers */
 unsigned long trace;
 /* bitmask of trace recursion */
 unsigned long trace_recursion;
#endif /* CONFIG_TRACING */
#ifdef CONFIG_CGROUP_MEM_RES_CTLR /* memcg uses this to do batch job */
 struct memcg_batch_info {
  int do_batch; /* incremented when batch uncharge started */
  struct mem_cgroup *memcg; /* target memcg of uncharge */
  unsigned long bytes;   /* uncharged usage */
  unsigned long memsw_bytes; /* uncharged mem+swap usage */
 } memcg_batch;
#endif
};